openstackgerrit | John Wu proposed openstack/rally: [Swift] Add objects context class https://review.openstack.org/179578 | 00:20 |
---|---|---|
*** PerfBeing has quit IRC | 00:34 | |
*** dmorita has joined #openstack-rally | 00:35 | |
*** PerfBeing has joined #openstack-rally | 00:35 | |
*** PerfBeing has quit IRC | 00:39 | |
*** exploreshaifali has quit IRC | 01:06 | |
*** rook has joined #openstack-rally | 01:08 | |
rook | boris-42 around by chance? | 01:08 |
*** PerfBeing has joined #openstack-rally | 01:10 | |
*** nkhare has quit IRC | 01:29 | |
*** PerfBeing has quit IRC | 01:57 | |
*** PerfBeing has joined #openstack-rally | 01:57 | |
*** PerfBeing has quit IRC | 02:02 | |
*** pradeep has joined #openstack-rally | 02:17 | |
*** baker has joined #openstack-rally | 02:34 | |
*** PerfBeing has joined #openstack-rally | 02:37 | |
*** baker_ has joined #openstack-rally | 02:43 | |
*** baker has quit IRC | 02:46 | |
*** coolsvap|afk is now known as coolsvap | 03:42 | |
*** baker_ has quit IRC | 03:48 | |
*** nkhare has joined #openstack-rally | 03:49 | |
*** pradeep has quit IRC | 03:53 | |
*** baker has joined #openstack-rally | 04:03 | |
openstackgerrit | John Wu proposed openstack/rally: [Swift] Add objects context class https://review.openstack.org/179578 | 04:23 |
*** baker has quit IRC | 04:26 | |
*** PerfBeing has quit IRC | 04:29 | |
*** PerfBeing has joined #openstack-rally | 04:30 | |
*** Vishal_ has joined #openstack-rally | 04:33 | |
*** PerfBeing has quit IRC | 04:35 | |
*** pradeep has joined #openstack-rally | 04:48 | |
*** abehl has joined #openstack-rally | 04:54 | |
*** abehl has quit IRC | 04:55 | |
*** abehl has joined #openstack-rally | 04:55 | |
*** yfried has quit IRC | 04:59 | |
*** turul_ has quit IRC | 05:07 | |
*** rdas has joined #openstack-rally | 05:10 | |
*** tfreger has joined #openstack-rally | 05:15 | |
*** kiran-r has joined #openstack-rally | 05:32 | |
*** nkhare_ has joined #openstack-rally | 05:40 | |
*** rdas has quit IRC | 05:46 | |
*** nkhare has quit IRC | 05:59 | |
*** yfried has joined #openstack-rally | 06:14 | |
*** neeti has joined #openstack-rally | 06:16 | |
*** kalyan has joined #openstack-rally | 06:31 | |
kalyan | Hi | 06:32 |
kalyan | i tried to use "periodic" in runner type | 06:32 |
kalyan | and it throws this error "There is no benchmark runner with type `periodic`" | 06:32 |
*** kiran-r has quit IRC | 06:35 | |
*** kevinbenton has quit IRC | 06:36 | |
*** rdas has joined #openstack-rally | 06:44 | |
*** gema_ is now known as gema | 06:54 | |
*** yingjun has joined #openstack-rally | 07:02 | |
*** e0ne has joined #openstack-rally | 07:16 | |
*** arxcruz has joined #openstack-rally | 07:28 | |
*** karimb has joined #openstack-rally | 07:37 | |
*** kiran-r has joined #openstack-rally | 08:10 | |
*** kiranr has joined #openstack-rally | 08:10 | |
*** kiranr has quit IRC | 08:10 | |
*** kiran-r is now known as kiranr | 08:11 | |
*** yingjun has quit IRC | 08:16 | |
*** amaretskiy has joined #openstack-rally | 08:17 | |
*** exploreshaifali has joined #openstack-rally | 08:18 | |
openstackgerrit | Mikhail Dubov proposed openstack/rally: Release 0.0.4 https://review.openstack.org/180498 | 08:19 |
*** yingjun has joined #openstack-rally | 08:21 | |
*** yingjun has quit IRC | 08:21 | |
*** kevinbenton has joined #openstack-rally | 08:32 | |
*** neeti has quit IRC | 08:39 | |
*** e0ne has quit IRC | 08:40 | |
*** neeti has joined #openstack-rally | 08:41 | |
*** agarciam has quit IRC | 09:07 | |
boris-42 | kalyan: replied in private | 09:31 |
boris-42 | rook: around | 09:31 |
boris-42 | rook: but seems like to late.. | 09:31 |
*** tosky has joined #openstack-rally | 09:38 | |
*** e0ne has joined #openstack-rally | 09:44 | |
*** e0ne is now known as e0ne_ | 09:44 | |
*** e0ne_ is now known as e0ne | 09:52 | |
openstackgerrit | Pavel Boldin proposed openstack/rally: Add command-dict option to specify command args https://review.openstack.org/177016 | 09:53 |
openstackgerrit | Pavel Boldin proposed openstack/rally: Introduce command-specifying dictionary https://review.openstack.org/177011 | 09:53 |
openstackgerrit | Pavel Boldin proposed openstack/rally: Introduce `valid_command' validator https://review.openstack.org/177012 | 09:53 |
openstackgerrit | Pavel Boldin proposed openstack/rally: Make `boot_runcommand_delete' accept command-dict https://review.openstack.org/177014 | 09:53 |
openstackgerrit | Pavel Boldin proposed openstack/rally: Add command-dict option to upload a local command https://review.openstack.org/177015 | 09:53 |
openstackgerrit | Pavel Boldin proposed openstack/rally: Fix `sshutils.put_file' to use sftp context https://review.openstack.org/182991 | 09:53 |
*** dmorita has quit IRC | 09:57 | |
*** cdent has joined #openstack-rally | 10:08 | |
*** aix has joined #openstack-rally | 10:20 | |
openstackgerrit | Merged openstack/rally: Fix security groups cleanup https://review.openstack.org/179428 | 10:21 |
openstackgerrit | Merged openstack/rally: Change default prefix in _generate_random_name() https://review.openstack.org/182732 | 10:22 |
*** kevinbenton has quit IRC | 10:25 | |
*** kevinbenton has joined #openstack-rally | 10:26 | |
*** aix has quit IRC | 10:42 | |
*** yfried is now known as yfried|afk | 10:42 | |
openstackgerrit | Alexander Maretskiy proposed openstack/rally: [Fuel] Add Fuel scenario `list_environments' https://review.openstack.org/178731 | 10:49 |
*** mwagner_lap has joined #openstack-rally | 11:05 | |
rook | boris-42 - Is there a change to Rally to use Sec-group id instead of the name? | 11:11 |
rook | boris-42: I seem to be bumping into this : Multiple security groups found matching 'default'. Use an ID to be more specific | 11:11 |
rook | with the latest code. | 11:11 |
*** e0ne is now known as e0ne_ | 11:12 | |
*** e0ne_ is now known as e0ne | 11:19 | |
*** yfried|afk is now known as yfried | 11:24 | |
*** kairat_kushaev has quit IRC | 11:24 | |
*** kairat_kushaev has joined #openstack-rally | 11:29 | |
boris-42 | rook: where that happened? | 11:31 |
rook | which scenario? | 11:31 |
rook | launching a bunch of VMs concurrently in a single tenant. | 11:32 |
rook | Nova.BootServer | 11:32 |
rook | i think was the scenario, lemme check | 11:32 |
rook | NovaServers.boot_server | 11:32 |
boris-42 | rook: please paste traces | 11:33 |
rook | boris-42: from ? Rally? | 11:33 |
boris-42 | rook: at least | 11:33 |
rook | https://gist.githubusercontent.com/jtaleric/21fa6e96251d9ac52e91/raw/23f89248d15ab2c73bc1c382f8d4e89e181432dc/gistfile1.txt | 11:35 |
rook | boris-42: ^ | 11:35 |
boris-42 | rook: give me a five minutes | 11:35 |
boris-42 | I am in middle of cutting 0.0.4 release | 11:35 |
openstackgerrit | Merged openstack/rally: Release 0.0.4 https://review.openstack.org/180498 | 11:35 |
rook | rgr | 11:36 |
*** fhubik has joined #openstack-rally | 11:39 | |
boris-42 | HAPPY 0.0.4 RELEASE | 11:42 |
rook | :D | 11:42 |
*** kairat_kushaev has quit IRC | 11:43 | |
boris-42 | rook: actually not fully yet | 11:43 |
rook | tease | 11:43 |
*** rdas has quit IRC | 11:49 | |
*** mwagner_lap has quit IRC | 11:58 | |
*** aix has joined #openstack-rally | 11:58 | |
pradeep | f | 12:04 |
*** exploreshaifali has quit IRC | 12:04 | |
*** e0ne is now known as e0ne_ | 12:05 | |
*** e0ne_ is now known as e0ne | 12:06 | |
boris-42 | rook: it's internal bug of your OpenStack cloud | 12:07 |
openstackgerrit | Mikhail Dubov proposed openstack/rally: [Spec] Enhance Rally API https://review.openstack.org/182245 | 12:13 |
*** dpaterson has joined #openstack-rally | 12:17 | |
openstackgerrit | Swapnil Kulkarni proposed openstack/rally: Add E126 pep8 rule https://review.openstack.org/183029 | 12:21 |
*** nkhare_ has quit IRC | 12:21 | |
*** kairat_kushaev has joined #openstack-rally | 12:22 | |
*** coolsvap is now known as coolsvap|afk | 12:22 | |
*** nkhare has joined #openstack-rally | 12:38 | |
boris-42 | rook: or maybe our bug | 12:39 |
*** yfried is now known as yfried|afk | 12:40 | |
*** yfried|afk is now known as yfried | 12:41 | |
vaidy | boris-42, | 12:45 |
vaidy | boris-42, ping | 12:45 |
vaidy | yfried, hi | 12:46 |
*** mwagner_lap has joined #openstack-rally | 12:46 | |
*** e0ne is now known as e0ne_ | 12:51 | |
*** e0ne_ is now known as e0ne | 12:53 | |
*** tfreger has quit IRC | 12:54 | |
*** yfried has quit IRC | 12:56 | |
*** pradeep has quit IRC | 13:00 | |
*** igormarnat is now known as imarnat | 13:14 | |
openstackgerrit | Mikhail Dubov proposed openstack/rally: [Docs] Existing users and other fixes https://review.openstack.org/183038 | 13:25 |
rook | boris-42: it is because rally doens't use a ID for sec groups | 13:26 |
*** Vishal_ has quit IRC | 13:26 | |
openstackgerrit | Mikhail Dubov proposed openstack/rally: [Docs] Existing users and other fixes https://review.openstack.org/183038 | 13:27 |
openstackgerrit | Andrey Kurilin proposed openstack/rally: Implement subunit_parser https://review.openstack.org/180144 | 13:29 |
boris-42 | rook: it is because you have few default groups | 13:29 |
rook | boris-42: my code cleans up sec groups before each rally run | 13:29 |
rook | to augment this issue i create more tenants and users. | 13:29 |
boris-42 | rook: https://github.com/openstack/rally/blob/master/rally/benchmark/scenarios/nova/utils.py#L116-L147 take a look | 13:30 |
boris-42 | rook: rally doesn't manage default groups at all | 13:30 |
boris-42 | rook: btw we fixed that issue (and rally cleans default groups now) | 13:30 |
rook | :) | 13:30 |
rook | good so i don't have to anymore | 13:30 |
rook | boris-42: something is creating def sec-groups when rally runs | 13:31 |
rook | maybe it is neutron | 13:31 |
boris-42 | rook: nova or nuetron | 13:32 |
boris-42 | rook: depending on what you are using | 13:32 |
rook | there is a known race condition in neutron which is fixed in kilo | 13:32 |
boris-42 | rook: yep I know | 13:32 |
rook | and rally exposes this pretty bad | 13:32 |
boris-42 | rook: I detected it | 13:32 |
boris-42 | rook: and I was helping fixing it=) | 13:32 |
*** PerfBeing has joined #openstack-rally | 13:33 | |
rook | boris-42: nice! | 13:35 |
rook | get it backported to juno would ya :) | 13:35 |
*** nkhare has quit IRC | 13:38 | |
rook | boris-42: ping - should kwargs["security_groups"] , be the ID or name? | 13:42 |
*** e0ne is now known as e0ne_ | 13:47 | |
openstackgerrit | Nikita Konovalov proposed openstack/rally: Wait for status https://review.openstack.org/172466 | 13:56 |
boris-42 | rook: hm | 14:04 |
rook | boris-42: seems to be a list of names :) | 14:04 |
boris-42 | rook: it should be list of names as far as I know | 14:04 |
rook | haha | 14:04 |
boris-42 | rook: yep | 14:04 |
rook | trial and error | 14:04 |
rook | it would be very nice to identify the kwargs and what they are expecting | 14:05 |
rook | not all of them | 14:05 |
boris-42 | rook: it's impossible | 14:05 |
rook | but what each function might use. | 14:05 |
boris-42 | rook: it depends on python clients commands that we are using | 14:05 |
boris-42 | rook: so you should take a look each time in specific python client | 14:06 |
rook | haha i have. | 14:06 |
rook | it would just be a Nice to have. | 14:06 |
*** e0ne_ is now known as e0ne | 14:06 | |
boris-42 | rook: so it will be nice if python clients are better and they will provide such info | 14:06 |
rook | preach | 14:06 |
boris-42 | rook: but it is impossible to organize such work in community | 14:06 |
boris-42 | rook: at least I am not able to lead such efforts=) | 14:07 |
boris-42 | I am to aggressive =) | 14:07 |
boris-42 | rook: retweet required https://twitter.com/PavlovicBoris/status/598838569793429505 =) | 14:07 |
rook | gah, even passing a a sec-group name and i still hit : RescheduledException: Build of instance ca97c8c4-e65d-453c-a120-37b47fd3959d was re-scheduled: Multiple security groups found matching 'default'. Use an ID to be more specific.\n" | 14:07 |
openstackgerrit | Roman Vasilets proposed openstack/rally: Add functional test for task samples https://review.openstack.org/181400 | 14:07 |
openstackgerrit | Roman Vasilets proposed openstack/rally: Fix nova/boot-lock-unlock-and-delete scenario flavor https://review.openstack.org/183053 | 14:07 |
boris-42 | rook: there is no way to fix this | 14:10 |
boris-42 | rook: except using "existing users" | 14:10 |
boris-42 | rook: which is possible now | 14:10 |
*** kiranr has quit IRC | 14:12 | |
openstackgerrit | Nikita Konovalov proposed openstack/rally: [Sahara] More debug logging https://review.openstack.org/170491 | 14:14 |
rook | boris-42: might give that a shot to avoid this. | 14:16 |
rook | boris-42: have a example of the existing user work | 14:17 |
*** baker has joined #openstack-rally | 14:19 | |
*** baker has quit IRC | 14:19 | |
*** baker has joined #openstack-rally | 14:19 | |
openstackgerrit | Roman Vasilets proposed openstack/rally: Add functional test for task samples https://review.openstack.org/181400 | 14:21 |
*** fhubik has quit IRC | 14:26 | |
*** fhubik has joined #openstack-rally | 14:26 | |
*** aswadr has joined #openstack-rally | 14:27 | |
*** imarnat is now known as igormarnat | 14:31 | |
*** stpierre has joined #openstack-rally | 14:32 | |
*** igormarnat is now known as imarnat | 14:34 | |
*** fhubik has quit IRC | 14:38 | |
*** fhubik has joined #openstack-rally | 14:39 | |
*** imarnat is now known as igormarnat | 14:40 | |
*** fhubik is now known as fhubik_afk | 14:41 | |
*** igormarnat is now known as imarnat | 14:43 | |
*** klindgren has joined #openstack-rally | 14:49 | |
openstackgerrit | Roman Vasilets proposed openstack/rally: Add functional test for task samples https://review.openstack.org/181400 | 14:50 |
*** imarnat is now known as igormarnat | 14:51 | |
*** fhubik_afk is now known as fhubik | 14:53 | |
*** igormarnat is now known as imarnat | 14:54 | |
*** fhubik has quit IRC | 14:58 | |
*** Vishal_ has joined #openstack-rally | 14:59 | |
*** imarnat is now known as igormarnat | 14:59 | |
*** baker has quit IRC | 15:04 | |
*** igormarnat is now known as imarnat | 15:04 | |
*** baker has joined #openstack-rally | 15:05 | |
Vishal_ | amaretskiy: Hi | 15:08 |
amaretskiy | Vishal_: hi | 15:08 |
Vishal_ | amaretskiy: Need your help for https://review.openstack.org/#/c/174251/ | 15:08 |
Vishal_ | amaretskiy: not able to decide what can be the persistent thing across iterations except "context" | 15:09 |
amaretskiy | Vishal_: simple answer: locate L35 at wrappers.network: cidr_incr = utils.RAMInt() | 15:14 |
*** arxcruz has quit IRC | 15:14 | |
amaretskiy | Vishal_: but your patch is interesting, I have to review it to give better comments | 15:14 |
Vishal_ | amaretskiy: sure | 15:14 |
amaretskiy | Vishal_: actually I got different problem with cidrs - how to avoid possible averlapping :) Your patch tries to do exactly opposite thing | 15:15 |
Vishal_ | amaretskiy: :) | 15:16 |
*** PerfBeing has quit IRC | 15:16 | |
*** PerfBeing has joined #openstack-rally | 15:16 | |
vaidy | boris-42, ping | 15:17 |
boris-42 | vaidy: pong | 15:18 |
vaidy | boris-42, first of nice to hear on the new release | 15:18 |
boris-42 | vaidy: thanks | 15:19 |
vaidy | boris-42, had a question on deployment create, is it mandatory that the json should have the admin user as mandatory entry to create a deployment? | 15:20 |
vaidy | boris-42, i see even the code is been designed the same way | 15:20 |
boris-42 | vaidy: for now yea | 15:20 |
boris-42 | vaidy: I am working on dropping this | 15:20 |
Vishal_ | amaretskiy: I feel there should be benchmarking done when overlapping Ip's are there in neutron...since i had seen different performance of neutron in case of overlapping and non-overlapping | 15:20 |
*** PerfBeing has quit IRC | 15:20 | |
vaidy | boris-42, ok | 15:21 |
boris-42 | vaidy: as first part we drop the requiremtns to create new users | 15:22 |
boris-42 | vaidy: so now you can use existing users in system to do the load | 15:22 |
boris-42 | vaidy: next step is to finish plugin base, move validation to it, and unify validation mechansim for context/scenarios/runners/sla | 15:22 |
openstackgerrit | Merged openstack/rally: Add E126 pep8 rule https://review.openstack.org/183029 | 15:23 |
vaidy | boris-42, ok | 15:24 |
*** baker has quit IRC | 15:24 | |
vaidy | boris-42, but makes i can understand the current architecture of the code | 15:24 |
boris-42 | vaidy: so current architecture of the code is in the middle of moving from generating load just with admin user -> to the abbility to generate load with what you have | 15:25 |
boris-42 | admin, admin + temp users create by rally, admin + existing users, existings users, or even just without openstack at all | 15:26 |
boris-42 | vaidy: ^ | 15:26 |
boris-42 | vaidy: as you can see on scenarios there is validator that ask do you need openstack and the rest stuff | 15:26 |
vaidy | boris-42, yeah | 15:26 |
vaidy | boris-42, I have not gone that far though but can co-relate what you are explaining | 15:27 |
*** neeti has quit IRC | 15:28 | |
boris-42 | vaidy: https://github.com/openstack/rally/blob/master/rally/benchmark/scenarios/nova/servers.py#L85 | 15:28 |
boris-42 | vaidy: ^ this | 15:28 |
openstackgerrit | Nikita Konovalov proposed openstack/rally: [Sahara] More debug logging https://review.openstack.org/170491 | 15:28 |
*** abehl has quit IRC | 15:31 | |
vaidy | boris-42, yeah | 15:36 |
vaidy | boris-42, thanks | 15:36 |
boris-42 | vaidy: at least now you can run rally on production like deployments=) | 15:36 |
*** baker has joined #openstack-rally | 15:37 | |
boris-42 | vaidy: becasue before it was just impossible | 15:37 |
vaidy | boris-42, yes we already are trying it | 15:37 |
vaidy | boris-42, but we are restricted with non admin access though | 15:37 |
vaidy | boris-42, this should solve that problem | 15:37 |
boris-42 | vaidy: so this will be implemented during liberty | 15:40 |
vaidy | boris-42, hmmm | 15:40 |
*** pradeep has joined #openstack-rally | 15:41 | |
vaidy | boris-42, ok | 15:41 |
Vishal_ | amaretskiy: The problem is the tenant and user for any particular iteration are chosen randomly...for the same tenant I have to generate a unique cidr so that within the tenant there is no overlapping...but for different tenant I want to use the same cidr ranges which were generated earlier so that they can overlap | 15:42 |
*** imarnat is now known as igormarnat | 15:48 | |
*** igormarnat is now known as imarnat | 15:51 | |
*** PerfBeing has joined #openstack-rally | 15:54 | |
*** MaxPC has joined #openstack-rally | 15:54 | |
*** imarnat is now known as igormarnat | 15:56 | |
*** igormarnat is now known as imarnat | 15:59 | |
*** yfried has joined #openstack-rally | 16:03 | |
openstackgerrit | Roman Vasilets proposed openstack/rally: Add functional test for task samples https://review.openstack.org/181400 | 16:03 |
*** asrangne has joined #openstack-rally | 16:05 | |
openstackgerrit | Chris St. Pierre proposed openstack/rally: Minor install script improvements https://review.openstack.org/182809 | 16:06 |
*** aswadr has quit IRC | 16:08 | |
*** abehl has joined #openstack-rally | 16:13 | |
*** e0ne is now known as e0ne_ | 16:18 | |
*** pradeep has quit IRC | 16:24 | |
*** e0ne_ is now known as e0ne | 16:25 | |
*** pradeep has joined #openstack-rally | 16:39 | |
*** abehl has quit IRC | 16:45 | |
*** abehl has joined #openstack-rally | 16:59 | |
rook | boris-42 - is there a example on using rally w/ existing users? | 16:59 |
*** amaretskiy has quit IRC | 17:03 | |
*** imarnat is now known as igormarnat | 17:05 | |
*** e0ne has quit IRC | 17:05 | |
boris-42 | rook: take a look at my blogpost | 17:07 |
boris-42 | rook: http://boris-42.me/rally-can-generate-load-with-passed-users-now/ | 17:07 |
rook | http://boris-42.me/rally-v0-0-4-whats-new/ ? | 17:07 |
rook | ah | 17:07 |
rook | ok | 17:07 |
boris-42 | rook: official documentation coming soon | 17:08 |
boris-42 | rook: actually patch is on review | 17:08 |
openstackgerrit | Roman Vasilets proposed openstack/rally: Add functional test for task samples https://review.openstack.org/181400 | 17:11 |
*** cdent has quit IRC | 17:17 | |
openstackgerrit | Merged openstack/rally: Add Cinder create_backup volume scenario https://review.openstack.org/167171 | 17:18 |
rook | boris-42: nice, running with that now. | 17:22 |
*** PerfBeing has quit IRC | 17:23 | |
*** PerfBeing has joined #openstack-rally | 17:23 | |
boris-42 | rook: the nice thing is that it won't leave bunch of undeleted resources | 17:26 |
boris-42 | rook: I mean when you use existing users (it's like a sandbox) | 17:27 |
*** PerfBeing has quit IRC | 17:28 | |
rook | boris-42: fyi rally task abort is broke. | 17:31 |
rook | known? | 17:31 |
*** Vishal_ has quit IRC | 17:31 | |
boris-42 | rook: it's not broken | 17:31 |
boris-42 | rook: it's not implmeneted | 17:31 |
rook | lol | 17:31 |
rook | ok | 17:32 |
boris-42 | rook: it says that it's not implmemented LOL | 17:32 |
* rook should of looked at the trace | 17:32 | |
rook | when return 1 i just assume "broke" | 17:32 |
rook | lol | 17:32 |
boris-42 | rook: eh end users=) | 17:33 |
rook | lol | 17:33 |
rook | it has been that sort of day | 17:33 |
boris-42 | rook: sooo | 17:33 |
boris-42 | rook: actually it's almost implmented | 17:33 |
boris-42 | кrook https://review.openstack.org/#/c/161636/ | 17:34 |
rook | boris-42: awesome... added myself as a reviewer thanks! | 17:35 |
boris-42 | rook: someday Rally will be awesome tool=) | 17:41 |
rook | boris-42: it is pretty helpful as it is now ;-) | 17:41 |
*** harlowja has quit IRC | 17:41 | |
*** harlowja has joined #openstack-rally | 17:42 | |
boris-42 | rook: https://review.openstack.org/#/c/147765/ | 17:42 |
boris-42 | rook: community=) | 17:42 |
rook | there are plenty of jerks in Open Source :) | 17:43 |
rook | boris-42 not saying one is better than the other... but Tempest has ruined OpenStack deployments in the past for me... It doens't do a good job at ALL when cleaning up after itself. | 17:45 |
boris-42 | rook: actually we are going to dicuss this during summit | 17:52 |
boris-42 | rook: so we have a lot of plans regarding to tuning cleanup stuff in rally | 17:52 |
rook | nice! | 17:52 |
boris-42 | rook: that is related to tempest as well | 17:52 |
boris-42 | rook: at least disaster cleanup part | 17:52 |
*** igormarnat is now known as imarnat | 17:53 | |
boris-42 | rook: begging is here https://review.openstack.org/#/c/172967/ | 17:53 |
*** e0ne has joined #openstack-rally | 17:53 | |
rook | rally does a much better job IMHO wrt Cleanup | 17:54 |
boris-42 | rook: so the future step are what to if server with rally and it's DB was destroyed in the middle of benchmarking?) | 18:06 |
*** PerfBeing has joined #openstack-rally | 18:06 | |
*** PerfBeing has quit IRC | 18:09 | |
*** PerfBeing has joined #openstack-rally | 18:09 | |
*** PerfBein_ has joined #openstack-rally | 18:12 | |
*** PerfBeing has quit IRC | 18:14 | |
*** exploreshaifali has joined #openstack-rally | 18:14 | |
*** PerfBein_ has quit IRC | 18:16 | |
*** georgem1 has joined #openstack-rally | 18:21 | |
openstackgerrit | Chris St. Pierre proposed openstack/rally: Add Nova scenario to boot multiple servers in one request https://review.openstack.org/176858 | 18:23 |
openstackgerrit | Chris St. Pierre proposed openstack/rally: Nova: Refactor utils._boot_servers to generate its own name prefix https://review.openstack.org/183152 | 18:23 |
*** tfreger has joined #openstack-rally | 18:30 | |
*** exploreshaifali has quit IRC | 18:30 | |
*** imarnat is now known as igormarnat | 18:33 | |
*** igormarnat is now known as imarnat | 18:35 | |
*** aix has quit IRC | 18:38 | |
*** botang has quit IRC | 18:40 | |
rook | hm, boris-42 what i find interesting... I tried with the singe existing user and hit the same issue iwth security groups | 18:42 |
*** botang has joined #openstack-rally | 18:42 | |
*** imarnat is now known as igormarnat | 18:42 | |
*** igormarnat is now known as imarnat | 18:42 | |
*** imarnat is now known as igormarnat | 18:43 | |
boris-42 | rook: yep | 18:46 |
boris-42 | rook: you should create user | 18:47 |
boris-42 | rook: boot single VM | 18:47 |
boris-42 | rook: remove it | 18:47 |
boris-42 | rook: after that you can benchmark | 18:47 |
rook | hm, the default sec-group exists without that. | 18:47 |
rook | | e586a999-3c59-42e3-b457-3b88a59a2226 | default | default | | 18:47 |
rook | is there something else that does? | 18:47 |
*** botang_ has joined #openstack-rally | 18:48 | |
boris-42 | rook: after creating user | 18:49 |
boris-42 | rook: you should create VM to create defaut sec group | 18:49 |
boris-42 | rook: that is actually the race | 18:49 |
boris-42 | rook: if you create them sinulatnously | 18:49 |
boris-42 | rook: it may create 2 default | 18:49 |
*** ajiang_ has joined #openstack-rally | 18:50 | |
*** botang has quit IRC | 18:50 | |
*** ajiang_ has quit IRC | 18:53 | |
*** ajiang has joined #openstack-rally | 18:54 | |
*** ajiang has quit IRC | 18:57 | |
*** ajiang has joined #openstack-rally | 18:58 | |
*** ajiang has left #openstack-rally | 19:01 | |
*** harlowja has quit IRC | 19:03 | |
*** harlowja has joined #openstack-rally | 19:05 | |
*** ajiang has joined #openstack-rally | 19:06 | |
*** abehl has quit IRC | 19:11 | |
georgem1 | I ran the task "NovaServers.boot_server" with "times": 10 and "concurrency": 10. After it completed successfully, I don't see any new instances | 19:20 |
*** PerfBeing has joined #openstack-rally | 19:21 | |
georgem1 | can somebody please explain me how it works…this is the task http://pastebin.com/m6E83BxQ | 19:22 |
georgem1 | I assumed it would create 3 temporary tenants with 2 users each and use them to start these VMs (10x10 ?), which should not be deleted at the end..also, what should happen with the temporary tenants and users? | 19:24 |
*** harlowja has quit IRC | 19:25 | |
*** harlowja has joined #openstack-rally | 19:26 | |
georgem1 | anybody? | 19:38 |
*** botang_ has quit IRC | 19:38 | |
*** botang has joined #openstack-rally | 19:38 | |
*** botang has quit IRC | 19:41 | |
boris-42 | georgem1: hm? | 19:44 |
boris-42 | georgem1: nope | 19:44 |
boris-42 | 10 times is total amount | 19:44 |
boris-42 | so it will create 10 workers each worker will create 1 vm | 19:44 |
boris-42 | after that generic cleanup mechanism will delete everything | 19:44 |
georgem1 | ok, where are the vms? | 19:44 |
boris-42 | they are cleaned up | 19:44 |
boris-42 | by Rally | 19:45 |
boris-42 | Rally don't leave any resources | 19:45 |
boris-42 | (or at least it shouldn't) | 19:45 |
georgem1 | even if I only ran boot | 19:45 |
boris-42 | cleanup is separated steps | 19:45 |
boris-42 | step* | 19:45 |
boris-42 | during the load it won't clean anything | 19:45 |
boris-42 | after load will be finished it will clean all VMs and all tenants/users | 19:46 |
georgem1 | so, what's the difference between the tasks "NovaServers.boot_server" and "NovaServers.boot_and_delete_server" ? | 19:46 |
boris-42 | load is different | 19:47 |
boris-42 | you are repeating N times creae or create-and-delete | 19:47 |
boris-42 | as a result in boot_server there will be more and more active VMs | 19:47 |
boris-42 | and in case of boot_and_delete the amount of active/in status of creation/deleteion will be the same as concurrency all the time | 19:48 |
boris-42 | georgem1: ^ | 19:48 |
georgem1 | I see, so create-and-delete deletes them right away, as opposed to just create that deletes at the end | 19:48 |
*** cdent has joined #openstack-rally | 19:49 | |
boris-42 | georgem1: yep after all iterations are run | 19:49 |
boris-42 | georgem1: so if you would like to get at the end 100 active VMs by creating 10 simultaneosuly | 19:49 |
boris-42 | times: 100, concurrency: 10 | 19:49 |
*** abehl has joined #openstack-rally | 19:49 | |
georgem1 | ok, actually, I want to test with 500, so it should be times: 500, concurrency 50 ? | 19:50 |
georgem1 | it will run 10 batches of 50 ? | 19:51 |
boris-42 | georgem1: there are not actually batches | 19:51 |
boris-42 | georgem1: there will be 50 workers that are simultaneously executing this stuff | 19:51 |
boris-42 | georgem1: scenario | 19:51 |
*** openstackgerrit has quit IRC | 19:52 | |
boris-42 | georgem1: unit it won't be run 500 times | 19:52 |
boris-42 | georgem1: depending on duration | 19:52 |
*** neeti has joined #openstack-rally | 19:52 | |
*** openstackgerrit has joined #openstack-rally | 19:52 | |
*** neeti has quit IRC | 19:52 | |
georgem1 | I see, but it's fair to say there will be 500 nova boot API calls in the end | 19:52 |
boris-42 | georgem1: it may happen that one worker will run 51 and another 49 | 19:52 |
boris-42 | georgem1: yep | 19:52 |
georgem1 | cool, thanks for the info | 19:53 |
*** cdent has quit IRC | 19:57 | |
georgem1 | boris-42: another question please..how does rally verify the instance started is really alive? | 20:01 |
boris-42 | georgem1: if it is boot_server than it will be just status in nova db | 20:01 |
boris-42 | georgem1: so if nova says that it works then it's ok | 20:02 |
georgem1 | boris-42: is there a task that also pings the instance's IP? | 20:02 |
*** tfreger has quit IRC | 20:04 | |
*** tfreger has joined #openstack-rally | 20:05 | |
*** abehl has quit IRC | 20:05 | |
*** asrangne has quit IRC | 20:09 | |
*** tfreger has quit IRC | 20:09 | |
*** asrangne has joined #openstack-rally | 20:10 | |
boris-42 | georgem1: so there is VMTask.boot_and_runcommand | 20:13 |
boris-42 | georgem1: it boot vm attach floating ip wait for ping wait for ssh and executes command inside vm collects output and remove VM | 20:13 |
georgem1 | boris-42:this sounds more accurate, I just wished it wouldn't use a floating IP as it's not going too well for large scale testing; it would probably run into arp issues, floating ip alloc/dealloc issues, etc; it would be nice if the first VM would be used as a jumpserver, but too complicated to implement.. | 20:17 |
*** abehl has joined #openstack-rally | 20:17 | |
*** igormarnat is now known as imarnat | 20:18 | |
boris-42 | georgem1: it can work with fixed ip | 20:21 |
boris-42 | georgem1: you need just to specify | 20:21 |
boris-42 | georgem1: to not use floatip | 20:21 |
boris-42 | georgem1: but as you understand you should see the fixed network from where do you run rally | 20:21 |
georgem1 | boris-42: how do I specify the user to run as and the network to attach to? I can run rally from an instance owned by a "rally" user in the "rally" tenant if need be | 20:24 |
*** baker has quit IRC | 20:30 | |
*** baker has joined #openstack-rally | 20:32 | |
boris-42 | georgem1: basically you can create user | 20:37 |
boris-42 | georgem1: and use it for generation load | 20:37 |
boris-42 | georgem1: or have rally on controllers | 20:37 |
boris-42 | georgem1: it really depends on your network configuration | 20:37 |
*** abehl has quit IRC | 20:42 | |
*** tfreger has joined #openstack-rally | 20:42 | |
*** georgem1 has quit IRC | 20:46 | |
*** tfreger has quit IRC | 20:49 | |
*** jaypipes has quit IRC | 20:53 | |
*** abehl has joined #openstack-rally | 20:53 | |
*** baker has quit IRC | 21:02 | |
*** imarnat is now known as igormarnat | 21:04 | |
*** PerfBein_ has joined #openstack-rally | 21:11 | |
*** botang has joined #openstack-rally | 21:11 | |
*** PerfBeing has quit IRC | 21:13 | |
*** PerfBein_ has quit IRC | 21:15 | |
*** openstackgerrit has quit IRC | 21:22 | |
*** openstackgerrit has joined #openstack-rally | 21:22 | |
*** baker has joined #openstack-rally | 21:23 | |
*** mwagner_lap has quit IRC | 21:25 | |
*** georgem1 has joined #openstack-rally | 21:33 | |
*** yfried is now known as yfried|afk | 21:35 | |
*** PerfBeing has joined #openstack-rally | 21:38 | |
*** e0ne has quit IRC | 21:39 | |
*** e0ne has joined #openstack-rally | 21:40 | |
openstackgerrit | Pavel Boldin proposed openstack/rally: Add command-dict option to specify command args https://review.openstack.org/177016 | 21:41 |
openstackgerrit | Pavel Boldin proposed openstack/rally: Fix `sshutils.put_file' to use sftp context https://review.openstack.org/182991 | 21:41 |
openstackgerrit | Pavel Boldin proposed openstack/rally: Introduce command-specifying dictionary https://review.openstack.org/177011 | 21:41 |
openstackgerrit | Pavel Boldin proposed openstack/rally: Introduce `valid_command' validator https://review.openstack.org/177012 | 21:41 |
openstackgerrit | Pavel Boldin proposed openstack/rally: Make `boot_runcommand_delete' accept command-dict https://review.openstack.org/177014 | 21:41 |
openstackgerrit | Pavel Boldin proposed openstack/rally: Add command-dict option to upload a local command https://review.openstack.org/177015 | 21:41 |
openstackgerrit | Pavel Boldin proposed openstack/rally: WIP command_customizer content https://review.openstack.org/183227 | 21:41 |
openstackgerrit | Pavel Boldin proposed openstack/rally: WIP: boot_runcommand_delete_custom_image https://review.openstack.org/183228 | 21:41 |
*** yfried|afk is now known as yfried | 21:46 | |
*** igormarnat is now known as imarnat | 21:48 | |
*** imarnat is now known as igormarnat | 21:49 | |
openstackgerrit | Pavel Boldin proposed openstack/rally: WIP: boot_runcommand_delete_custom_image https://review.openstack.org/183228 | 21:50 |
*** PerfBeing has quit IRC | 21:53 | |
*** igormarnat is now known as imarnat | 21:53 | |
*** PerfBeing has joined #openstack-rally | 21:53 | |
*** abehl has quit IRC | 21:54 | |
*** yfried is now known as yfried|afk | 21:57 | |
*** PerfBeing has quit IRC | 21:58 | |
*** e0ne has quit IRC | 22:01 | |
*** imarnat is now known as igormarnat | 22:04 | |
*** igormarnat is now known as imarnat | 22:06 | |
*** imarnat is now known as igormarnat | 22:06 | |
*** igormarnat is now known as imarnat | 22:08 | |
*** georgem1 has quit IRC | 22:08 | |
*** imarnat is now known as igormarnat | 22:10 | |
*** igormarnat is now known as imarnat | 22:19 | |
*** nkhare has joined #openstack-rally | 22:22 | |
*** baker has quit IRC | 22:30 | |
*** tosky has quit IRC | 22:35 | |
*** botang_ has joined #openstack-rally | 22:35 | |
*** botang has quit IRC | 22:38 | |
*** imarnat is now known as igormarnat | 22:45 | |
*** PerfBeing has joined #openstack-rally | 22:47 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/rally: Updated from global requirements https://review.openstack.org/182843 | 22:49 |
*** igormarnat is now known as imarnat | 22:59 | |
*** oanufriev has joined #openstack-rally | 23:06 | |
*** davideagnello has quit IRC | 23:08 | |
*** asrangne__ has joined #openstack-rally | 23:08 | |
*** imarnat is now known as igormarnat | 23:10 | |
*** asrangne has quit IRC | 23:11 | |
*** igormarnat is now known as imarnat | 23:11 | |
*** davideagnello has joined #openstack-rally | 23:11 | |
*** imarnat is now known as igormarnat | 23:11 | |
*** nkhare has quit IRC | 23:17 | |
openstackgerrit | Mikhail Dubov proposed openstack/rally: [Docs] Existing users and other fixes https://review.openstack.org/183038 | 23:33 |
*** igormarnat is now known as imarnat | 23:50 | |
*** imarnat is now known as igormarnat | 23:52 | |
*** igormarnat is now known as imarnat | 23:54 | |
*** imarnat is now known as igormarnat | 23:55 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!