Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2015-02-09

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

All times shown according to UTC.

Time Nick Message
00:55 psd_ joined #openstack-rally
01:08 rook joined #openstack-rally
01:46 zhiyan joined #openstack-rally
02:19 psd_ joined #openstack-rally
02:28 mwagner_lap joined #openstack-rally
02:57 rook joined #openstack-rally
03:16 nkhare joined #openstack-rally
03:52 clfh joined #openstack-rally
04:39 pboldin joined #openstack-rally
04:45 rook joined #openstack-rally
04:46 pboldin1 joined #openstack-rally
04:47 openstackgerrit Pavel Boldin proposed stackforge/rally: Add FIO disk benchmark for VMs  https://review.openstack.org/152388
04:47 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the generic script-based benchmark for the VMs  https://review.openstack.org/141671
04:47 openstackgerrit Pavel Boldin proposed stackforge/rally: Add network-aware VM boot/delete utils  https://review.openstack.org/151384
04:47 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the context benchmark_image  https://review.openstack.org/138466
04:47 openstackgerrit Pavel Boldin proposed stackforge/rally: Boot server w/ predefined internal network  https://review.openstack.org/152387
04:47 openstackgerrit Pavel Boldin proposed stackforge/rally: Fix get_floating_ip lookup for {"id": net_id"}  https://review.openstack.org/152766
04:58 psd__ joined #openstack-rally
05:06 pboldin joined #openstack-rally
05:37 Mehrnaz joined #openstack-rally
05:44 pboldin1 joined #openstack-rally
06:04 pboldin joined #openstack-rally
06:26 boris-42 joined #openstack-rally
06:49 yfried joined #openstack-rally
06:53 yfried_ joined #openstack-rally
07:07 Mehrnaz Hi every body.I still have problem in installing Rally. It couldn't be install on debian wheezy! :(
07:12 neeti joined #openstack-rally
07:35 psd joined #openstack-rally
07:42 yfried_ pboldin: ping
07:43 pboldin yfried_: pong
07:43 Miouge joined #openstack-rally
07:43 yfried_ pboldin: https://review.openstack.org/#/c/149771/16/rally/benchmark/context/keypair.py,cm.
07:43 pboldin yfried_: well?
07:43 yfried_ pboldin: L77 - could you help me with understanding the cleanup please?
07:44 pboldin yfried_: a resource manager is used to cleanup all the users keypairs
07:44 yfried_ pboldin: where did the keypairs register under "nova.keypairs"?
07:45 pboldin yfried_: they don't. the resource manager is using a SeekAndDestroy strategy
07:45 yfried_ pboldin: can you point me to the code?
07:45 pboldin yfried_: sure, wait a second
07:45 yfried_ will it destroy all keypairs that belong to the users in the list?
07:46 pboldin yfried_: it should
07:46 pboldin yfried_: https://github.com/stackforge/rally/tree/master/rally/benchmark/context/cleanup
07:46 nmagnezi joined #openstack-rally
07:49 yfried_ pboldin: so this assumes no keypairs were created for these users outside the keypair context?
07:50 pboldin yfried_: yes. and this is incompatible with the 'existing users'
07:50 pboldin yfried_: but it yet for boris-42 to decide the approach here
07:50 yfried_ pboldin: that was going to be my next q
07:50 yfried_ pboldin: tnx. will continue to review
07:50 boris-42 Mehrnaz: hi there
07:51 Mehrnaz boris-42: Hi boris
07:51 boris-42 Mehrnaz: I filed bug
07:51 boris-42 Mehrnaz: https://bugs.launchpad.net/rally/+bug/1419449
07:51 Mehrnaz boris-42: thanks, now I should use debian jessie for installing Rally?
07:51 boris-42 Mehrnaz: so we will need some time to fix it
07:52 boris-42 Mehrnaz: I didn't fix it...
07:52 boris-42 Mehrnaz: I just field..
07:52 Mehrnaz This problem was for debian wheezy, not jessie
07:52 boris-42 Mehrnaz: oh
07:52 boris-42 Mehrnaz: ok
07:53 Mehrnaz Did I tell you jessie? If i did, I'm so sorry. that was my fault.
07:53 boris-42 Mehrnaz: fixed bug description
07:53 boris-42 Mehrnaz: no worries fixed*
07:53 Mehrnaz boris-42: thanks.
07:53 rdas joined #openstack-rally
07:53 boris-42 Mehrnaz: so could you for now use ubuntu or centos?
07:53 boris-42 Mehrnaz: and we will try to figure out why debian doesn't work..
07:55 e0ne joined #openstack-rally
07:57 boris-42 yfried_: pboldin guys it is
07:57 boris-42 yfried_: pboldin comable with "existing users"
07:57 pboldin boris-42: oh. how it is so?
07:57 boris-42 yfried_: pboldin the thing is that cleanup stuff is still in progress
07:57 boris-42 yfried_: pboldin https://review.openstack.org/#/c/139643/
07:58 pboldin yfried_: oh, so you are not going to delete anything w/o the prefix?
07:58 boris-42 pboldin: yfried_ yep I am going to use prefixes
07:58 boris-42 pboldin: yfried_ this is the only way
07:58 pboldin boris-42: yep
07:58 boris-42 pboldin: so it will be .list() + filtering by prefixes
07:59 yfried_ boris-42: why not register resourced for cleanup?
08:00 boris-42 yfried_: you mean store somewhere resources?
08:00 yfried_ boris-42: instead of sticking to naming, have each resource register its UUID in a "to clean" list
08:00 boris-42 yfried_: it is very bad approach, unsafe for production clouds
08:00 yfried_ boris-42: why?
08:01 boris-42 yfried_: and as well it is bad for performance of rally it self
08:01 openstackgerrit Pavel Boldin proposed stackforge/rally: Add FIO disk benchmark for VMs  https://review.openstack.org/152388
08:01 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the generic script-based benchmark for the VMs  https://review.openstack.org/141671
08:01 openstackgerrit Pavel Boldin proposed stackforge/rally: Add floating-IP aware VM boot/delete utils  https://review.openstack.org/151384
08:01 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the context benchmark_image  https://review.openstack.org/138466
08:01 yfried_ boris-42: you tag what should be destroy at the end of test
08:01 openstackgerrit Pavel Boldin proposed stackforge/rally: Boot server w/ predefined internal network  https://review.openstack.org/152387
08:01 boris-42 yfried_: plus it is hard and complicated=)
08:01 boris-42 yfried_: so all scenarios in rally should be refactored
08:01 boris-42 yfried_: to somehow register all resources
08:01 yfried_ boris-42: ok
08:01 boris-42 yfried_: this resources should be stored in persistance storge
08:02 boris-42 yfried_: otherwise if rally process fail => leak of resources
08:02 boris-42 yfried_: next thing is that storing all UUIDs before creating resources will require a lot of CPU/RAM/Network...
08:02 yfried_ boris-42: if rally process fail, you're gonna have leakage anyway. unless you are counting on the next rally run to remove those
08:02 boris-42 yfried_ it is our goal
08:02 yfried_ boris-42: ack
08:03 arxcruz joined #openstack-rally
08:03 boris-42 yfried_: to have method
08:03 boris-42 "disaster cleanup"
08:03 boris-42 that can be run even with new instance of rally on new server without old DB
08:03 boris-42 yfried_: ^
08:03 boris-42 yfried_: it's quite big and hard task but I see how it can be implemented..
08:05 yfried_ boris-42: https://review.openstack.org/#/c/148187/ https://review.openstack.org/#/c/148637/
08:05 yfried_ boris-42: could you please review these 2?
08:06 yfried_ boris-42: the second one is a hell to rebase each time
08:06 boris-42 yfried_: uh oh
08:06 boris-42 yfried_: atomic mixin
08:07 boris-42 yfried_: ok I will take a look today
08:07 yfried_ pboldin: could you please explain to me https://review.openstack.org/#/c/149771/16/rally/benchmark/context/secgroup.py,cm
08:07 yfried_ pboldin: L113
08:08 boris-42 yfried_: please don't use '
08:08 boris-42 yfried_: one guy almost died replacing ' to "
08:08 pboldin yfried_: what is your question about it?
08:08 yfried_ boris-42: I spent ~1hr yesterday trying to hunt the code for all of it
08:08 boris-42 yfried_: ?
08:08 boris-42 yfried_: there is rally job for that
08:08 yfried_ pboldin: when do we open ssh for admin and why?
08:09 pboldin yfried_: this is to be used in the following patches.
08:09 boris-42 yfried_: https://review.openstack.org/#/c/148637/6 Mirantis Rally CI pep8
08:09 boris-42 yfried_: http://185.8.56.88/logs/WDdQjTRhyOqiLkcZ/pep8/epep8.txt.gz
08:09 boris-42 yfried_: there are few that were before your patch
08:09 boris-42 yfried_:  but you are introducing a lot
08:09 yfried_ pboldin: haven't seen those patches yet, but will they be working in admin tenant and with admin user to create and ssh to VMs?
08:10 boris-42 yfried_: the second thing I would prefer another name
08:10 pboldin yfried_: if there is 'admin' user, yes
08:10 boris-42 yfried_: self.atomic_action() instead of self.timer()
08:10 boris-42 yfried_: actually I will left comment in review
08:11 yfried_ boris-42: please sync with amaretskiy  since he argues for the other way :)
08:11 yfried_ pboldin: I'm looking at the titles for the next patches, none of them states tests that run in admin tenant
08:12 fhubik joined #openstack-rally
08:12 pboldin yfried_: https://review.openstack.org/#/c/138466/
08:12 fhubik joined #openstack-rally
08:15 yfried_ pboldin: I really don't like that. it opens the gate for abusement and errors that involve admin tenant
08:16 yfried_ pboldin: couldn't you create the instance in a regular tenant and use admin to make it public?
08:16 pboldin yfried_: well, users are created using the admin
08:16 pboldin yfried_: yeah, this is a variant too.
08:16 yfried_ pboldin: I don't like allowing rally to create resource IN admin tenant
08:16 pboldin yfried_: i spoke it to boris-42 but we have not discussed it.
08:17 yfried_ well I'm going to look into it more when I get to it, but I'm thinking about -1
08:17 pboldin yfried_: ok
08:18 pboldin yfried_: i've lost the hope to merge them anyway :D
08:18 boris-42 yfried_: he always argue in differnet way
08:19 boris-42 yfried_: he has own way
08:19 boris-42 yfried_: but what I think is that if person see in one place @atomic_action in another self.timer() he will never think that this about the same
08:19 boris-42 yfried_: and this is bad in root
08:20 yfried_ boris-42: I agree. but I'm following his guide for this whole patch, so I would like him to sign up on that change
08:21 boris-42 yfried_: your patch is not so simple as you may think
08:22 yfried_ boris-42: I never said it was simple. when did you read that?
08:22 boris-42 yfried_: =)
08:22 boris-42 yfried_:  I mean it is not simple for me
08:22 yfried_ boris-42: well, baiscall it's moving class from scenario to common and then changing all the refernces to it
08:23 boris-42 yfried_: yep but for now we have only one class that uses it
08:23 boris-42 yfried_: it's base.Scenario
08:24 yfried_ boris-42: we'll be able to use it in context and get timers for creation and deletion of stuff as well
08:24 openstackgerrit Koert van der Veer proposed stackforge/rally: Add a (configurable) set of roles to users  https://review.openstack.org/153595
08:25 neeti joined #openstack-rally
08:35 rook joined #openstack-rally
08:35 boris-42 yfried_: ok
08:36 boris-42 yfried_: but I dislike self.timer name
08:36 boris-42 yfried_: so let's do next
08:36 boris-42 yfried_: I will left comments, I will ask amaretsky to review it
08:36 yfried_ boris-42: ack
08:37 boris-42 yfried_: we will align everything
08:37 yfried_ boris-42: sound reasonable
08:37 boris-42 yfried_: and only after that you'll change code. so you won't waste time refactoring it over and over
08:40 andreykurilin_ joined #openstack-rally
08:40 fhubik joined #openstack-rally
08:41 fhubik_afk joined #openstack-rally
08:48 psd boris-42, Hey!
09:09 pboldin yfried_: ping. can you please advice me how can I get the Internet access inside the VMs?
09:12 boris-42 psd: hey hey
09:13 psd boris-42, whatsup!
09:14 boris-42 psd: heh working=)
09:16 psd boris-42, well, regarding https://review.openstack.org/#/c/113177/
09:16 psd boris-42, I am supposed to make changes to the already written code, right?
09:20 boris-42 psd: yep fetch that code rebase it and fix
09:20 psd boris-42, Alrighty :)
09:22 yfried_ pboldin: what do you mean?
09:22 yfried_ pboldin: actually be able to access google?
09:22 pboldin yfried_: my VM is unable to access external network.
09:22 pboldin yfried_: yeah, for instance
09:23 yfried_ pboldin: rally tests shouldn't do it
09:23 boris-42 yfried_: why they shouldn't do it?
09:23 boris-42 yfried_: oh?
09:24 yfried_ boris-42: test should be able to run on every cloud
09:25 yfried_ boris-42: clouds aren't always connected to outside
09:25 yfried_ boris-42: and you shouldn't allow VMs to pull/send traffic outside the cloud anyway for security reasons
09:25 yfried_ boris-42: definitly not during test
09:26 pboldin yfried_: but what if our test requires outside connection for instance to download an application?
09:26 yfried_ boris-42: the furthest you sould reach during a test is the external network GW/DNS which can be retrieved from subnet details
09:27 yfried_ pboldin: you sould prepare the image ahead of test run
09:28 boris-42 yfried_: that approac actually sux=)
09:28 boris-42 just imagine
09:28 yfried_ pboldin: you could provide an external script/tool to prepare an image, but the actual test should hold data DL from implicit sources
09:28 boris-42 you have at least 200 becnh tools
09:29 pboldin boris-42: yfried_: well, i'm going off to sleep right now :-/
09:29 boris-42 storing somewhere 200 images is bad
09:29 msdubov joined #openstack-rally
09:29 boris-42 yfried_: btw why it has secure concerns?
09:29 yfried_ pboldin: how can you sleep when your cloud is downloading unkown and an approved data :)?
09:29 boris-42 yfried_: what about end users?
09:30 boris-42 yfried_: that are running VMs with internet?
09:30 yfried_ boris-42: what do you mean?
09:31 yfried_ boris-42: I need to review the actual code.
09:33 boris-42 yfried_: I mean I don't see any security concerns
09:34 boris-42 yfried_: public cloud you are giving to the end user VM with internet connectivity
09:34 boris-42 yfried_: it's the same thing that we need
09:34 boris-42 just VM with internet
09:34 yfried_ boris-42: aren't you being billed according to your bandwidth usage?
09:35 boris-42 yfried_: so not sure that I understand how it is related
09:35 yfried_ boris-42: your test is suddently generating unkown but large amount of data DL
09:35 boris-42 yfried_: to the benchmark
09:35 boris-42 yfried_: not large
09:36 boris-42 yfried_: size to download and install tool that you need once
09:36 yfried_ boris-42: I need to look at the actual DL code before I can continue this discussion
09:36 boris-42 yfried_: sure
09:36 boris-42 yfried_: I mean just to make it clear
09:37 anshul joined #openstack-rally
09:37 boris-42 yfried_: we are making only one image * not on every run of VMTask.boot_runncomand_delete
09:39 amaretskiy joined #openstack-rally
09:45 yfried_ boris-42: "
09:45 yfried_ Otherwise create one image
09:45 yfried_ per user and tenant
09:45 yfried_ "
09:46 boris-42 yfried_: per tenant*
09:46 boris-42 yfried_: and actually for those kind of tests you can sepify only one tenant
09:46 boris-42 yfried_: so it will create one image as well
09:48 yfried_ boris-42: if you accidentally/intentionally use multiple tenants, you'll get multiple snapshots
09:49 psd boris-42,  I am trying to add a new function to rally/benchmark/scenarios/neutron/utils.py, for removing interface router, but I am not able to find how should I call that from a neutron client
09:51 yfried_ boris-42: paboldin: for package installation you can use "cloud-init" instead of "installation script"
09:52 boris-42 yfried_: yep if you use multple tenants without admin passed to rally it will create N snapshots
09:52 boris-42 yfried_: and?
09:52 boris-42 yfried_: if you put to big load to Rally it will DDoS openstack
09:53 yfried_ boris-42: I'm not sure what we are talking about anymore ... :)
09:53 boris-42 yfried_: so yep it's hard to make benchmarking tool that doesn't allow you to shout your legs
09:53 boris-42 yfried_: I mean if end user decide to shoot his legs
09:53 boris-42 yfried_: we won't be able to stop him =)
10:00 e0ne joined #openstack-rally
10:07 rook joined #openstack-rally
10:07 psd boris-42, I am able to locate json api docs(http://developer.openstack.org/api-ref-guides/bk-api-ref.pdf)  but I am not able to find the docs for the python api
10:07 psd json calls*
10:08 boris-42 psd: the fastest way that I know is next
10:08 boris-42 psd: just open python client code
10:08 boris-42 psd: https://github.com/openstack/python-neutronclient/tree/master/neutronclient/neutron/v2_0
10:09 e0ne joined #openstack-rally
10:11 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add new script for gate jobs  https://review.openstack.org/144633
10:22 psd boris-42, Thanks a lot
10:24 e0ne joined #openstack-rally
11:17 openstackgerrit Merged stackforge/rally: Implement abort in scenario runners  https://review.openstack.org/151678
11:31 openstackgerrit Vitaly Gusev proposed stackforge/rally: Add rally tests for Ceilometer  https://review.openstack.org/153994
11:36 openstackgerrit Mikhail Dubov proposed stackforge/rally: Stop scenario execution on SLA failure  https://review.openstack.org/152459
11:37 openstackgerrit Mikhail Dubov proposed stackforge/rally: Stop scenario execution on SLA failure  https://review.openstack.org/152459
11:43 openstackgerrit Roman Vasilets proposed stackforge/rally: Support scenarios for updating neutron quotas  https://review.openstack.org/92794
11:51 e0ne joined #openstack-rally
11:54 cdent joined #openstack-rally
11:54 rook joined #openstack-rally
12:02 Miouge joined #openstack-rally
12:28 nkhare joined #openstack-rally
12:29 openstackgerrit Yair Fried proposed stackforge/rally: Adds support for multiple tempest sources  https://review.openstack.org/148187
12:34 openstackgerrit Yair Fried proposed stackforge/rally: Adds support for multiple tempest sources  https://review.openstack.org/148187
12:34 yfried_ andreykurilin: ^ are these unittests ok?
12:36 tellesnobrega boris-42, hi, does rally work properly with keystone v3?
12:40 andreykurilin yfried_: you forgot to remove L177:) https://review.openstack.org/#/c/148187/9/tests/unit/verification/test_tempest.py
12:40 openstackgerrit Ivan Kolodyazhny proposed stackforge/rally: Implement create_and_upload_volume_to_image scenario for Cinder  https://review.openstack.org/152670
12:41 boris-42 tellesnobrega: yep it works
12:42 tellesnobrega boris-42, if i set endpoint type to private it users keystone:5000?
12:43 openstackgerrit Yair Fried proposed stackforge/rally: Adds support for multiple tempest sources  https://review.openstack.org/148187
12:43 yfried_ andreykurilin: please check again
12:43 boris-42 tellesnobrega: you should use this deployment config https://github.com/stackforge/rally/blob/master/samples/deployments/existing-keystone-v3.json
12:43 tellesnobrega thanks
12:44 boris-42 tellesnobrega: so you should specify project instead of tenant info
12:44 boris-42 tellesnobrega: and rally will use keystone v3 instead of v2
12:44 tellesnobrega great
12:50 andreykurilin yfried_: http://logs.openstack.org/87/148187/7/check/rally-coverage/95a0bd2/cover/rally_verification_tempest_tempest.html _is_git_repo,  _move_contents_to_subdir and base_repo are weakly tested:(
12:50 andreykurilin yfried_: sorry, old coverage)
12:51 andreykurilin yfried_: should wait new results
12:52 yfried_ andreykurilin: well, it's a shutil,move call. don't see how/why we should test it
12:53 yfried_ andreykurilin: also, don't see how to test it without creating an actual old env
12:57 andreykurilin yfried_: we can mock shutil, so we will be able to check correct args and kwargs which transmitted there.
12:57 arxcruz joined #openstack-rally
13:10 rook joined #openstack-rally
13:19 nmagnezi joined #openstack-rally
13:24 tellesnobrega boris-42, when I'm running a task, rally tries to connect to keystone:35357 but I didnt create this endpoint, is there a way that i can force rally to use keystone:5000
13:31 openstackgerrit Prabhjyot Singh Sodhi proposed stackforge/rally: Adds deleting routers scenarios for Neutron:  https://review.openstack.org/113177
13:32 boris-42 amaretskiy: could you help tellesnobrega  ^
13:32 yfried_ andreykurilin: could you show me an example please?
13:32 boris-42 tellesnobrega: the reason why it goest to 35357 is because it's admin port by default
13:33 andreykurilin yfried_: sure, one moment
13:33 boris-42 tellesnobrega: and rally requires admin for now
13:34 psd boris-42, I have added a new patch, check it when you find time :) (https://review.openstack.org/#/c/113177/)
13:35 amaretskiy boris-42 ok
13:35 tellesnobrega boris-42, hum, so i need to add this endpoint to my keystone, thanks
13:35 amaretskiy tellesnobrega hi
13:35 tellesnobrega amaretskiy, hi
13:35 boris-42 psd: thank you
13:35 boris-42 tellesnobrega: ya if you add new endpoint it will help
13:35 psd boris-42, :D
13:36 amaretskiy tellesnobrega the solution is deploying rally with samples/deployments/existing-with-given-endpoint.json
13:36 boris-42 amaretskiy: I am not sure that that will help but we can try
13:36 boris-42 amaretskiy: because 5000 seems like a not admin port
13:37 tellesnobrega amaretskiy, i will give it a try
13:38 amaretskiy tellesnobrega please give us (via paste.openstack.org) the output of `rally deployment config' (you can shadow login/password)
13:38 tellesnobrega ok
13:41 tellesnobrega amaretskiy, this is what i got now, after putting endpoint property
13:41 tellesnobrega http://paste.openstack.org/show/169935/
13:41 andreykurilin yfried_: sorry for delay. please, look at https://github.com/stackforge/rally/blob/master/tests/unit/verification/test_tempest.py#L187 and https://github.com/stackforge/rally/blob/master/tests/unit/verification/test_tempest.py#L199-L201
13:41 oanufriev joined #openstack-rally
13:43 amaretskiy tellesnobrega if issue still occurs, give us the rally error message
13:43 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano benchmarks  https://review.openstack.org/137661
13:43 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano base for murano benchmarks  https://review.openstack.org/137650
13:44 tellesnobrega ok
13:44 tellesnobrega amaretskiy, thanks
13:50 Miouge joined #openstack-rally
13:57 tellesnobrega amaretskiy, hey, i'm getting this error now, http://paste.openstack.org/show/169953/
13:59 amaretskiy tellesnobrega have you run rally in verbose mode? (rally -v {command})
13:59 tellesnobrega doing now
13:59 amaretskiy need more details about the problem
14:00 tellesnobrega amaretskiy, the problem is with my keystone instalation, tried to fetch users and got the same Resource not found
14:00 amaretskiy ok
14:05 e0ne joined #openstack-rally
14:11 tellesnobrega amaretskiy, here is the full log, http://paste.openstack.org/show/169958/
14:11 tellesnobrega the problem is it cant create new users
14:11 rediskin joined #openstack-rally
14:13 tellesnobrega amaretskiy, do you have any clue why?
14:14 rediskin hiyo
14:14 amaretskiy tellesnobrega please give the output of keystone catalog for identity service
14:14 amaretskiy 1) $ . ~/.rally/openrc
14:14 amaretskiy 2) $ keystone catalog --service identity
14:15 rediskin who have issues with debian + install-rally.sh?
14:15 jaypipes joined #openstack-rally
14:15 tellesnobrega http://paste.openstack.org/show/169959/
14:15 tellesnobrega amaretskiy, i set up rally with v3
14:16 amaretskiy ok, is there a catalog data for identity service?
14:18 tellesnobrega what do you mean by catalog data?
14:21 amaretskiy tellesnobrega http://paste.openstack.org/show/169968/
14:22 tellesnobrega yeah, http://paste.openstack.org/show/169959/
14:23 tellesnobrega amaretskiy, ^
14:25 amaretskiy tellesnobrega: now, can you create a user via `keystone user-create' cli command ?
14:32 tellesnobrega amaretskiy, no
14:32 amaretskiy what is the error message ?
14:32 amaretskiy from keystone cli
14:34 tellesnobrega just a second
14:36 openstackgerrit Yair Fried proposed stackforge/rally: Adds support for multiple tempest sources  https://review.openstack.org/148187
14:40 tellesnobrega amaretskiy, i'm getting now an authorization error, but i had that solved before. not sure what i did wrong now
14:41 amaretskiy what is the type of your cloud (devstack, fuel) ?
14:44 tellesnobrega amaretskiy, its just a keystone
14:45 tellesnobrega amaretskiy, it actually two keystones, behind a HA proxy
14:45 tellesnobrega using and distributed db
14:46 tellesnobrega i need to test keystone performance in this scenario. i heard that fuel can deploy this, but i havent used it
14:47 yfried_ andreykurilin: how is this https://review.openstack.org/#/c/148187/?
14:49 yfried_ boris-42: andreykurilin: amaretskiy: how can I execute mirantis CI pep8 locally?
14:50 amaretskiy tellesnobrega I think that the first step is ensure that we can create users via keystone client (cli is enough)
14:50 amaretskiy if this work - then we can start debuggung rally
14:50 openstackgerrit Yair Fried proposed stackforge/rally: Adds support for multiple tempest sources  https://review.openstack.org/148187
14:50 amaretskiy if not - we should probably find the issue on cloud-side
14:51 amaretskiy s/debuggung/debugging/
14:51 tellesnobrega ok
14:51 andreykurilin yfried_: one moment
14:54 andreykurilin yfried_: pep8 job on Mirantis CI differs only by one module: flake8-double-quotes(Flake8 lint to forbide single quotes.)
14:55 andreykurilin redixin/rediskin: can you share info about how flake8-double-quotes is enabled on our CI?)
14:56 tellesnobrega amaretskiy, so here is where i'm at. If i set the deployment with v3 i dont get the credentials with openrc, if I set if v2, i get the catalog and credentials, but i cant do any actions related to users, for some reason its not working. (this is my deploy problem for sure)
14:56 andreykurilin yfried_: your patch looks good to me, but let's wait vote from jenkins:)
14:57 yfried_ amaretskiy: could you please explain https://review.openstack.org/#/c/148637/6/tests/unit/common/test_utils.py,cm
14:58 amaretskiy tellesnobrega put `--debug' to rally command and give the part of output around the error
14:58 yfried_ amaretskiy: ^ how is the new test class different from the existing one?
14:59 rook joined #openstack-rally
14:59 amaretskiy yfraid_: you have tests for AtomicAction
14:59 amaretskiy yfried_: so where are tests for AtomicMixin ?
14:59 amaretskiy also, there is a missed test
15:00 amaretskiy for atomic_action_timer
15:00 amaretskiy that is a module-level function
15:00 amaretskiy wrapper
15:02 tellesnobrega amaretskiy, i'm using keystone v3 now, http://paste.openstack.org/show/169988/
15:02 tellesnobrega it creates one user, but only one
15:06 yfried_ amaretskiy: I'm not sure what kind of tests should be there
15:06 rediskin andreykurilin, "pip install flake8-double-quotes" thats all
15:06 * yfried_ sux at unittests
15:06 rediskin andreykurilin, flake8 seems automatically uses all installed extensions
15:07 andreykurilin yfried_: ^
15:07 amaretskiy yfried_ we need proper unittests for existing and added functions/classes, so my comment looks reasonable
15:10 yfried_ amaretskiy: I don't argue with that. I want to know which tests you are looking for
15:10 exploreshaifali joined #openstack-rally
15:11 amaretskiy yfried_ the rule is well-known: we need at least one test_* method for each function or method
15:12 yfried_ amaretskiy: well, I didn't know it...
15:12 amaretskiy yfried_ I can add a comment with a skeleton for tests
15:13 amaretskiy with all test methods names
15:13 yfried_ amaretskiy: I'd appreciate that.
15:13 amaretskiy ok
15:21 amaretskiy tellesnobrega: look at line 42 of your http://paste.openstack.org/show/169988/
15:22 amaretskiy there is an empty list of roles returned
15:23 openstackgerrit Anastasia Kuznetsova proposed stackforge/rally: Fix issues in Mistral code  https://review.openstack.org/154076
15:26 amaretskiy tellesnobrega: I think we should find out why there is an empty roles list
15:35 amaretskiy yfried_: I've posted a comment about tests
15:52 andreykurilin yfried_: here?
15:59 andreykurilin yfried_: please, skip failed test in https://review.openstack.org/#/c/148187/ like https://github.com/stackforge/rally/blob/master/tests/unit/verification/test_tempest.py#L128
16:00 rook joined #openstack-rally
16:05 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano benchmarks  https://review.openstack.org/137661
16:05 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano base for murano benchmarks  https://review.openstack.org/137650
16:06 yfried_ andreykurilin: ack
16:06 openstackgerrit Yair Fried proposed stackforge/rally: Adds support for multiple tempest sources  https://review.openstack.org/148187
16:07 yfried_ andreykurilin: what if someone actually tries to use rally with py26
16:08 andreykurilin yfried_: py26 is not supported by tempest and rally will not run verification on py26 - https://github.com/stackforge/rally/blob/master/rally/verification/tempest/tempest.py#L265-L273
16:09 andreykurilin yfried_: https://github.com/openstack/tempest/blob/master/setup.cfg#L18
16:09 yfried_ andreykurilin: how convenient
16:09 andreykurilin :)
16:11 andreykurilin yfried_: so we can use all benefits from py27 in "rally verify" in backword incompatible way(without support oldest python version)
16:11 andreykurilin yfried_: I suppose, we should check launch of tempest in py3 env
16:18 msdubov joined #openstack-rally
16:20 msdubov joined #openstack-rally
16:29 amaretskiy joined #openstack-rally
16:32 jjmb joined #openstack-rally
16:33 tellesnobrega amaretskiy, hum, will do
16:41 rook joined #openstack-rally
16:43 fhubik joined #openstack-rally
16:52 tellesnobrega amaretskiy, you got it. It was the missing role
16:52 amaretskiy nice :)
16:52 tellesnobrega thanks a lot
16:58 openstackgerrit Mikhail Dubov proposed stackforge/rally: Stop scenario execution on SLA failure  https://review.openstack.org/152459
17:22 tellesnobrega amaretskiy, is this something i should be worried about http://paste.openstack.org/show/170069/
17:24 amaretskiy tellesnobrega this error looks something like a bug - can you report this at launchpad? (however we need a way to reproduce that)
17:24 tellesnobrega sure
17:24 tellesnobrega i will write my setup so you guys can try to reproduce it
17:26 amaretskiy tellesnobrega thank you!
17:27 tellesnobrega no problem
17:34 pboldin joined #openstack-rally
17:37 tellesnobrega amaretskiy, https://bugs.launchpad.net/rally/+bug/1419910
17:38 amaretskiy tellesnobrega please add the task json
17:38 tellesnobrega sure
17:38 amaretskiy we need as much info for reproducing the issue as possible
17:38 amaretskiy thanks!
18:02 tellesnobrega no problem
18:04 harlowja joined #openstack-rally
18:07 psd joined #openstack-rally
18:40 meteorfox joined #openstack-rally
18:45 yfried_ joined #openstack-rally
18:55 psd_ joined #openstack-rally
19:29 exploreshaifali joined #openstack-rally
19:30 e0ne joined #openstack-rally
19:38 andreykurilin_ joined #openstack-rally
19:41 openstackgerrit joined #openstack-rally
19:46 psd__ joined #openstack-rally
21:07 ilbot3 joined #openstack-rally
21:07 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
21:14 psd__ joined #openstack-rally
21:16 dmellado joined #openstack-rally
21:19 nkhare joined #openstack-rally
21:22 boris-42 tellesnobrega: are you around?
21:22 boris-42 tellesnobrega: I was really busy whole day
21:23 boris-42 tellesnobrega: so that message is from Keystone service and it's due some issues of configuration of keystone
21:23 boris-42 (We are creating quite rapidly users)
21:24 meteorfox joined #openstack-rally
21:30 tellesnobrega boris-42, hum, i see
21:30 tellesnobrega i changed my setup and it is working now
21:31 boris-42 tellesnobrega: heh
21:31 tellesnobrega i had a distributed db using galera
21:31 boris-42 tellesnobrega: oh oh
21:31 tellesnobrega now i'm using a single mysql and its fine
21:31 boris-42 tellesnobrega: that thing can fail
21:32 boris-42 =)
21:40 psd__ joined #openstack-rally
22:00 dmellado joined #openstack-rally
22:07 psd|away joined #openstack-rally
22:38 ilbot3 joined #openstack-rally
22:38 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
22:52 dmellado joined #openstack-rally
22:56 clfh joined #openstack-rally
22:56 dmellado_ joined #openstack-rally
23:03 andreykurilin_ joined #openstack-rally
23:32 openstackgerrit Merged stackforge/rally: Adds support for multiple tempest sources  https://review.openstack.org/148187
23:38 dmellado joined #openstack-rally
23:39 openstackgerrit Merged stackforge/rally: Fix issues in Mistral code  https://review.openstack.org/154076

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