Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2014-11-14

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

All times shown according to UTC.

Time Nick Message
00:36 dmorita joined #openstack-rally
00:57 pboldin joined #openstack-rally
01:00 pboldin left #openstack-rally
01:09 Viswanath joined #openstack-rally
01:37 marcoemorais joined #openstack-rally
01:45 chandankumar joined #openstack-rally
01:51 rdas joined #openstack-rally
01:53 lordd joined #openstack-rally
01:56 openstackgerrit Merged stackforge/rally: Sahara Cluster error state handling  https://review.openstack.org/125311
02:04 openstackgerrit Merged stackforge/rally: Updated from global requirements  https://review.openstack.org/134386
02:13 aswadr joined #openstack-rally
02:30 openstackgerrit Boris Pavlovic proposed stackforge/rally: Refactor generic cleanup mechanism part 3  https://review.openstack.org/129060
02:37 lordd joined #openstack-rally
02:42 Poornima joined #openstack-rally
02:45 erkules_ joined #openstack-rally
02:46 yaguang joined #openstack-rally
03:04 lordd joined #openstack-rally
03:09 lordd joined #openstack-rally
03:29 lordd joined #openstack-rally
03:30 chandankumar joined #openstack-rally
03:33 nkhare joined #openstack-rally
03:46 sponomarenko joined #openstack-rally
03:47 coolsvap joined #openstack-rally
03:47 jseutter joined #openstack-rally
03:49 chandan_kumar joined #openstack-rally
03:54 lordd joined #openstack-rally
03:55 chandankumar joined #openstack-rally
04:13 AlexF joined #openstack-rally
04:24 lordd joined #openstack-rally
05:26 lordd joined #openstack-rally
05:40 chandankumar joined #openstack-rally
05:41 yaguang joined #openstack-rally
05:47 AlexF joined #openstack-rally
06:04 marcoemorais joined #openstack-rally
06:05 marcoemorais1 joined #openstack-rally
06:06 fandi joined #openstack-rally
06:09 rdas joined #openstack-rally
06:28 lordd joined #openstack-rally
06:44 neeti joined #openstack-rally
07:08 k4n0 joined #openstack-rally
07:29 lordd joined #openstack-rally
08:10 fandi joined #openstack-rally
08:18 chandankumar joined #openstack-rally
08:30 lordd joined #openstack-rally
08:35 chandankumar joined #openstack-rally
08:54 lordd joined #openstack-rally
08:55 oanufriev joined #openstack-rally
09:02 amaretskiy joined #openstack-rally
09:25 nkhare joined #openstack-rally
09:37 aswadr joined #openstack-rally
10:02 vkmc joined #openstack-rally
10:06 msdubov joined #openstack-rally
10:29 msdubov_ joined #openstack-rally
10:30 cdent joined #openstack-rally
10:35 nkhare joined #openstack-rally
10:43 rediskin joined #openstack-rally
10:48 lordd joined #openstack-rally
11:17 rdas joined #openstack-rally
11:29 lordd joined #openstack-rally
12:06 paboldin joined #openstack-rally
12:07 chandankumar joined #openstack-rally
12:52 pboros joined #openstack-rally
13:04 lordd joined #openstack-rally
13:05 nkhare joined #openstack-rally
13:43 openstackgerrit Boris Pavlovic proposed stackforge/rally: Refactor generic cleanup mechanism part 3  https://review.openstack.org/129060
13:46 lordd joined #openstack-rally
13:47 AlexF joined #openstack-rally
13:51 yaguang joined #openstack-rally
14:06 yaguang joined #openstack-rally
14:28 boris-42 rediskin andreykurilin could you take a look a tmy patch ^
14:29 andreykurilin boris-42, sure:)
14:30 berserker joined #openstack-rally
14:30 boris-42 berserker: heh nice nick name lol=)
14:30 berserker thnx boris
14:30 berserker I have a question regarding, pre-created tenants and users
14:31 berserker you might be the right one :)
14:31 boris-42 berserker: so I am working on that lol
14:31 berserker does Rally support the use of pre-created tenants and users for benchmarks?
14:32 boris-42 berserker:  you can hack for now rally
14:32 berserker right, I saw a comment from you
14:32 boris-42 berserker: it's not hard to hack, but proper solution takes some amount of time
14:32 berserker ok
14:33 boris-42 berserker: only thing that you should change is here
14:33 boris-42 berserker:  https://github.com/stackforge/rally/blob/master/rally/benchmark/context/users.py#L230-L258
14:33 boris-42 berserker:  instead of creating users in setup() return already existing
14:33 boris-42 berserker: and remove all code from cleanup
14:33 boris-42 berserker:  that is all
14:34 berserker boris-42: ok. where are the credentials taken from then?
14:35 boris-42 berserker:  so "it's hardcore" approach
14:35 boris-42 berserker:  as I said it's quite straight =)
14:36 berserker boris-42: that it is. but still I wonder, what credentials is the benchmark using then?
14:36 lordd joined #openstack-rally
14:36 berserker boris-42: or in other words, how do I provide the credentials for the benchmark then?
14:37 boris-42 berserker: so you have users
14:37 boris-42 berserker: that already existing you need bunch of stuff
14:37 boris-42 berserker:  auth_url (keystone url), username, password, teanntmae
14:38 msdubov_ joined #openstack-rally
14:38 boris-42 berserker: this code creates users https://github.com/stackforge/rally/blob/master/rally/benchmark/context/users.py#L181-L191
14:39 boris-42 berserker: so you need instead of creating them, just pass already existing
14:39 ntrueblood joined #openstack-rally
14:39 berserker boris-42: ok cool. and I suppose same for tenants
14:40 boris-42 berserker: yep
14:40 boris-42 berserker: and everything else will work
14:40 berserker boris-42: thanks alot. I will give it a try. Looking forward for your implementation
14:40 boris-42 berserker: in future you'll be able to pass users via ""rally deployment create"
14:41 boris-42 berserker: ya it's not hard to implement it, but I need to address validation of task for that
15:00 ntrueblood joined #openstack-rally
15:33 msdubov_ joined #openstack-rally
15:33 mcaulfie joined #openstack-rally
15:41 AlexF joined #openstack-rally
16:42 msdubov_ joined #openstack-rally
16:55 marcoemorais joined #openstack-rally
16:55 fandi joined #openstack-rally
17:04 jlk joined #openstack-rally
17:04 jlk Hey all, I'm new to rally trying to use it on a private cloud. Having some trouble with my deployment, I added my auth url which has a specific port, and when using rally in debug mode it appears to be using a different port than I specified. Could anybody give me a hand in figuring out what's going on?
17:05 lordd joined #openstack-rally
17:08 jlk looks like there is an admin_port setting
17:12 mwagner_lap joined #openstack-rally
17:33 mcaulfie_ joined #openstack-rally
17:35 harlowja joined #openstack-rally
17:37 harlowja joined #openstack-rally
17:46 leeantho joined #openstack-rally
18:05 marcoemorais joined #openstack-rally
18:06 marcoemorais joined #openstack-rally
18:16 lordd joined #openstack-rally
18:23 mcaulfie joined #openstack-rally
18:26 echoingumesh joined #openstack-rally
18:26 lordd joined #openstack-rally
18:32 boris-42 Jlk ya there is admin_port setting
18:33 jlk it was unfortunate that it didn't discover the admin port by the auth URL, but that's ffine.
18:33 jlk now ti figure out why it can't add users
18:33 boris-42 Jlk hm interesting
18:34 boris-42 Jkl what version of openstack are u testing?
18:35 marcoemorais joined #openstack-rally
18:35 jlk Havana
18:35 lordd joined #openstack-rally
18:38 boris-42 Jlk could u paste exception?
18:38 jlk doing a deploy check looks like it has trouble with OS-KSADM/services
18:38 jlk a 404 failure
18:41 jlk perhaps this cloud isn't running the openstack ksadm extension
18:51 boris-42 Jlk interesting
18:51 boris-42 Jlk actually rally uses latest pyton clients
18:51 boris-42 Jlk that can make as well sense
19:02 marcoemorais joined #openstack-rally
19:04 marcoemorais joined #openstack-rally
19:07 marcoemorais joined #openstack-rally
19:07 marcoemorais joined #openstack-rally
19:13 jlk boris-42: so you think that the client code is too new? We can use our old client code, but we set some ENV vars, like COMPUTE_API_VERSION=1.1 and NOVA_VERSION=1.1
19:14 jlk if I set those in the deployment config, will rally consume them?
19:15 boris-42 Nope it won't :(
19:15 boris-42 Jfk so actually we started working on rally during havana
19:16 boris-42 And there were no braking changes as far ad i remember
19:16 boris-42 Soo not sure why it doesn't work..
20:08 vkmc boris-42, hey dud :)
20:08 vkmc boris-42, could you reproduce the Zaqar's cleanup bug again?
20:18 boris-42 vkmc: hhmmm
20:18 boris-42 vkmc:  just run on any patch "check exeperimental"
20:18 boris-42 vkmc:  lemme do that
20:18 boris-42 vkmc:  hi there btw=)
20:19 boris-42 vkmc:  https://review.openstack.org/#/c/129060/ okay this patch
20:19 boris-42 vkmc:  so let's wait for a while=)
20:19 vkmc k :)
20:19 vkmc thanks
20:19 vkmc I was checking the code and its not what I think it was
20:33 lordd joined #openstack-rally
20:33 marcoemorais joined #openstack-rally
20:38 vkmc c'moooooooooooooon gate
20:38 vkmc >.>
20:57 boris-42 vkmc:  ?)
20:57 vkmc looks like a connection error
20:57 vkmc http://logs.openstack.org/60/129060/40/experimental/gate-rally-dsvm-zaqar-rally/643d4a8/console.html
20:58 boris-42 vkmc: something went wrong lol
20:58 boris-42 vkmc: ya it happens =)
20:58 boris-42 vkmc: lemme recheck it =)
20:58 vkmc :) of course yeah, thx
20:59 boris-42 vkmc:  =)
21:15 lordd joined #openstack-rally
21:16 vkmc boris-42, look at that... :P
21:16 vkmc https://review.openstack.org/#/c/129060/
21:16 vkmc http://logs.openstack.org/60/129060/40/experimental/gate-rally-dsvm-zaqar-rally/a507139/console.html
21:16 boris-42 vkmc: vkmc ya
21:16 boris-42 vkmc: there are some warnings lol
21:17 boris-42 vkmc:  1 ERROR rally.benchmark.context.cleanup.manager [-] The header X-PROJECT-ID was missing
21:17 vkmc AH ok
21:17 vkmc :p
21:17 vkmc you are making me work
21:17 vkmc hahaha
21:17 boris-42 vkmc: ahaha=)
21:18 boris-42 vkmc: https://github.com/stackforge/rally/blob/master/rally/benchmark/context/cleanup/resources.py#L265-L269
21:18 vkmc I see now the problem, thx
21:18 boris-42 vkmc: so this is the code of new generic cleanup stuff
21:18 boris-42 vkmc: list() command fails -)
21:18 vkmc boris-42, I see yeah
21:19 vkmc just in case... there is a typo there https://github.com/stackforge/rally/blob/master/rally/benchmark/context/cleanup/resources.py#L249
21:19 vkmc I'll add the project-id header to the queues call
21:21 boris-42 цруку црку?)
21:22 boris-42 where where?)
21:22 boris-42 vkmc: I mean typo ^
21:22 boris-42 ah I see
21:22 vkmc lool
21:22 boris-42 celiometer instead of ceilometer
21:22 boris-42 lol
21:22 * vkmc learns russian
21:22 boris-42 lol
21:22 vkmc yeah
21:22 boris-42 vkmc: dont' translate dat!
21:23 boris-42 vkmc: NOO!
21:23 boris-42 =)
21:23 * vkmc closes the translator site
21:23 vkmc ok :o
21:23 vkmc haha
21:24 boris-42 -)
21:24 vkmc soooo the typo is in the function signature, that's why I pointed out
21:24 * vkmc is not nitpicky
21:25 boris-42 vkmc: hm could you elaborate where exactly?)
21:25 * boris-42 I am nitpicky) totally nerd=)
21:27 vkmc oh sorry, I misread
21:27 vkmc its the class only
21:27 boris-42 ya=)
21:27 boris-42 vkmc: class doesn't make any sense )
21:27 lordd joined #openstack-rally
21:27 vkmc yup
21:28 boris-42 vkmc:  but you can be a hero =)
21:28 boris-42 vkmc:  and fix that=)
21:29 vkmc boris-42, a hero for a typo fix?
21:29 vkmc I could report it and let a newcomer take it
21:29 vkmc that would be even better
21:29 boris-42 vkmc: lol =)
21:30 boris-42 vkmc: yep or just fix with zaqar stuff=)
21:30 vkmc sounds like you want the second one :p
21:31 vkmc so lets do that
21:45 boris-42 vkmc: ya the second thing is better lol=)
21:46 boris-42 vkmc: I think it may be bug in zaqar client?
21:46 marcoemorais joined #openstack-rally
21:46 vkmc boris-42, no.. its ok
21:46 boris-42 vkmc:  hm hm?)
21:46 boris-42 vkmc:  okay=)
21:47 vkmc boris-42, the thing is that the project-id header is always needed, for most of the requests
21:47 boris-42 vkmc:  yep including creating of queue
21:47 vkmc boris-42, I'm trying to get the context and the project-id with that
21:47 boris-42 vkmc: that actually works!
21:47 boris-42 vkmc:  but why create queue works? it requres as well project-id=)
21:47 vkmc boris-42, creating queues doesn't need the project-id header :) it should IMO
21:47 boris-42 vkmc: okay seems like I am not an zaqar expert=)
21:48 * boris-42 vkmc: http://www.youtube.com/watch?v=B7MIJP90biM
21:48 boris-42 =)
21:48 vkmc boris-42, no haha, truth is that we should control for the project-id header for queues creation
21:49 * vkmc clicks
21:53 lordd joined #openstack-rally
21:54 boris-42 vkmc: did you see original "expert movie" ?)
21:56 vkmc boris-42, I haven't :)
21:56 boris-42 vkmc:  ahh soo
21:57 boris-42 vkmc: http://www.youtube.com/watch?v=BKorP55Aqvg
21:57 boris-42 =)
21:57 vkmc ohhhhhh that one I did!
21:58 vkmc haha the first one is the sequel, so cool
21:58 boris-42 vkmc: yyayaa
21:58 boris-42 =)
22:05 marcoemorais1 joined #openstack-rally
22:06 marcoemorais joined #openstack-rally
22:23 jlk joined #openstack-rally
22:26 openstackgerrit Victoria Martínez de la Cruz proposed stackforge/rally: Fix Zaqar's cleanup  https://review.openstack.org/134661
22:30 boris-42 flwang1 hi there
22:30 flwang1 boris-42: hey
22:30 boris-42 flwang1: sooo any news?)
22:30 flwang1 boris-42: no good progress, we're blocked on some env issue.
22:31 flwang1 boris-42: btw, now devstack can support SSL, so it would be nice if you can enable it for Rally
22:31 boris-42 flwang1: btw do we need to separated jobs?
22:31 boris-42 2*
22:31 AlexF joined #openstack-rally
22:34 flwang1 boris-42: i don't think so
22:34 flwang1 boris-42: but it depends
22:35 boris-42 flwang1: okay we will try to cover this as well
22:35 flwang1 boris-42: I believe no public cloud want to run openstack without SSL
22:35 flwang1 boris-42: but for private cloud, it depends
22:36 flwang1 boris-42: FWIW, it would be awesome if Rally can do some test against SSL
22:36 boris-42 flwang1: actually it can
22:36 flwang1 boris-42: btw, I saw there are some tasks for tempest
22:36 flwang1 so what is for?
22:36 boris-42 flwang1:  I am not sure that you are facing rally issue*
22:36 flwang1 test the performance of Tempest?
22:37 boris-42 flwang1: nope
22:37 boris-42 flwang1:  generate load using tempest scenarios
22:37 boris-42 flwang1: don't run that on production!
22:37 boris-42 =)
22:37 flwang1 boris-42: maybe Rally works great, but our env is not good.
22:37 flwang1 I'm still working on that
22:37 flwang1 will bug you as long as some issues come out :)
22:37 AlexF joined #openstack-rally
22:39 boris-42 flwang1: so but in any case we will test ssl as well
22:39 boris-42 flwang1: I have some ideas about how to make it
22:39 flwang1 boris-42: for now, we're trying to use tempest scenario test to do some 'smoke' test. that means, for example, we're going to launch a new openstack region, then we want to use scenario test to verify/confirm everything is ok
22:39 AlexF joined #openstack-rally
22:39 flwang1 and then we want to use Rally to confirm the performance is acceptable
22:39 boris-42 flwang1:  btw did you see "rally verify" command?)
22:39 flwang1 that's our use case
22:40 flwang1 boris-42: yep, but I haven't try that. May I get some info from Rally PTL? ;)
22:41 boris-42 flwang1:  hm so there is actually tutorial
22:42 boris-42 flwang1:  https://www.mirantis.com/blog/rally-openstack-tempest-testing-made-simpler/
22:42 flwang1 I think I just need to run 'rally verify start'
22:42 boris-42 flwang1:  ya
22:42 boris-42 flwang1:  and it will do all stuff lol
22:42 Viswanath joined #openstack-rally
22:42 boris-42 you can specify set by "--set image"
22:42 flwang1 boris-42: can it make sure the SSL work?
22:43 flwang1 boris-42: I mean if I enabled SSL in Rally, can Rally make sure enable SSL for tempest?
22:43 boris-42 flwang1: ohhh not sure=)
22:43 flwang1 boris-42: :(
22:43 boris-42 flwang1:  but you can pass to "start"
22:43 boris-42 flwang1:  any tempest.conf
22:43 openstackgerrit Victoria Martínez de la Cruz proposed stackforge/rally: Fix typo in Ceilometer's class in the cleanup  https://review.openstack.org/134664
22:43 boris-42 if you have it
22:43 flwang1 I'm running
22:44 flwang1 boris-42: btw, what's the test suite Tempest will run?
22:44 flwang1 all?
22:44 boris-42 flwang1: if you don't specify any set it will run "--set smoke"
22:45 boris-42 flwang1:  if you want all then you need to specify --set full
22:45 flwang1 so "rally verify start --set scenario" should also work, is it?
22:46 boris-42 flwang1: actually --regex
22:46 boris-42 flwang1:  will work in such case
22:46 boris-42 flwang1: if you want to run something specific not sure about --set scenario*
22:46 flwang1 boris-42: ok, sounds interesting. I love this feature
22:47 flwang1 I want to kiss the developer of this feature
22:47 marcoemorais1 joined #openstack-rally
22:48 boris-42 flwang1:  there is a lot of them lol
22:48 boris-42 flwang1:  btw results are stored in Rally.DB
22:48 flwang1 boris-42: cool
22:48 boris-42 flwang1: so you can get results in HTML, JSON or something else
22:48 boris-42 flwang1: + you can compare 2 runs*
22:48 flwang1 boris-42: I think you guys are doing amazing stuff than xxpest
22:49 flwang1 though I think xxpest is great as well :D
22:49 flwang1 i'm kidding
22:50 boris-42 flwang1: =)
22:50 vkmc xxpest with steroids
22:52 openstackgerrit Mikhail Dubov proposed stackforge/rally: Further improvements in rally info  https://review.openstack.org/134666
22:54 boris-42 vkmc: hm not sure that I get your question=)
22:55 vkmc boris-42, I don't remember exactly where are the endpoint attributes setted... but I remember that tenant_id wasn't there
22:55 vkmc boris-42, probably because it wasn't needed :p
22:55 boris-42 vkmc: ya it was only tenant_name
22:55 vkmc boris-42, so well, I can add it
22:55 boris-42 vkmc:  https://github.com/stackforge/rally/blob/master/rally/objects/endpoint.py#L30
22:56 vkmc boris-42, and set it during the initialization of the Zaqar client
22:56 boris-42 vkmc:  can you use name instead of ID?
22:56 vkmc hmm.. don't think so
22:56 boris-42 vkmc:  so we can translate it =)
22:56 vkmc there yeah :)
22:56 flwang1 boris-42: is there any way to run all tasks?
22:56 flwang1 boris-42: under the sample path
22:57 openstackgerrit Mikhail Dubov proposed stackforge/rally: Further improvements in rally info  https://review.openstack.org/134666
22:57 boris-42 flwang1: so the only way is to put them in one file
22:57 boris-42 flwang1: something like this https://github.com/stackforge/rally/blob/master/rally-scenarios/rally.yaml
22:57 boris-42 vkmc: soo
22:57 boris-42 vkmc: https://github.com/stackforge/rally/blob/master/rally/osclients.py#L268 you have here active keystone
22:57 boris-42 vkmc:  I belive there is somewhere inisde tenant_id
22:57 boris-42 vkmc:  lemme find it
22:59 flwang1 boris-42: and then what the command looks like?
22:59 vkmc boris-42, oh you are right
22:59 vkmc brb
23:00 boris-42 vkmc: so
23:00 boris-42 vkmc:  kc.tenant_id should return what you need
23:00 boris-42 flwang1: hm just "rally task start <that_file>"
23:01 boris-42 flwang1:  it will run all those benchmarks one by one
23:02 flwang1 boris-42: cool
23:03 boris-42 flwang1:  so task -> set of benhcmarks
23:03 flwang1 boris-42: and again, most of the user want to leverage Rally's default tasks, so it would nice if we can make sure all of them work
23:03 boris-42 flwang1: all commands rally task * works with set
23:03 boris-42 flwang1: they work
23:03 boris-42 flwang1:  more or less we are testing them
23:03 flwang1 boris-42: good
23:03 boris-42 flwang1:  you can help to write functional test for them if you would like
23:04 boris-42 flwang1: cause now there is only unit test
23:08 flwang1 boris-42: I will if i can get more bandwidth :)
23:10 boris-42 flwang1: when you find some time just ping me I'll help you
23:35 rediskin left #openstack-rally
23:38 marcoemorais joined #openstack-rally
23:47 openstackgerrit Merged stackforge/rally: Fix typo in Ceilometer's class in the cleanup  https://review.openstack.org/134664
23:57 marcoemorais joined #openstack-rally

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