Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2015-01-14

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

All times shown according to UTC.

Time Nick Message
00:08 rediskin joined #openstack-rally
00:13 yingjun joined #openstack-rally
00:17 openstackgerrit OpenStack Proposal Bot proposed stackforge/rally: Updated from global requirements  https://review.openstack.org/146713
00:41 jjmb joined #openstack-rally
00:46 jjmb1 joined #openstack-rally
00:48 zhiyan joined #openstack-rally
00:52 msdubov_ joined #openstack-rally
00:58 jjmb joined #openstack-rally
01:45 jjmb1 joined #openstack-rally
02:05 yaguang joined #openstack-rally
02:06 zhiyan joined #openstack-rally
02:41 shakamunyi joined #openstack-rally
02:41 mwagner_lap joined #openstack-rally
02:48 shakamunyi joined #openstack-rally
02:57 zhiyan joined #openstack-rally
04:40 flwang joined #openstack-rally
04:52 yaguang joined #openstack-rally
05:22 psd joined #openstack-rally
05:49 jjmb joined #openstack-rally
06:00 yfried__ joined #openstack-rally
06:02 tfreger joined #openstack-rally
06:03 yfried joined #openstack-rally
06:50 yingjun joined #openstack-rally
07:11 zhiyan joined #openstack-rally
07:25 rvasilets joined #openstack-rally
07:26 rdas joined #openstack-rally
07:46 jjmb1 joined #openstack-rally
07:54 nmagnezi joined #openstack-rally
07:56 prmtl joined #openstack-rally
08:19 openstackgerrit Prabhjyot Singh Sodhi proposed stackforge/rally: Adding documentation  https://review.openstack.org/144835
08:19 psd joined #openstack-rally
08:31 arxcruz joined #openstack-rally
08:34 yingjun_ joined #openstack-rally
08:45 oanufriev joined #openstack-rally
08:57 yingjun joined #openstack-rally
09:09 psd joined #openstack-rally
09:25 openstackgerrit Nikita Konovalov proposed stackforge/rally: Neutron network context support  https://review.openstack.org/143671
09:26 amaretskiy joined #openstack-rally
09:35 openstackgerrit Yuriy Nesenenko proposed stackforge/rally: Added list_snapshots scenario for Cinder test  https://review.openstack.org/146580
09:40 yingjun joined #openstack-rally
09:41 dmorita joined #openstack-rally
09:41 prmtl joined #openstack-rally
09:45 openstackgerrit Alexander Maretskiy proposed stackforge/rally: Improvements for scenario VMTasks.boot_runcommand_delete  https://review.openstack.org/144337
09:49 exploreshaifali joined #openstack-rally
09:54 rvasilets Hi. When I see in log "2015-01-14 04:16:37.649 9283 INFO rally.benchmark.context.users [-] Task af4dd9bd-2e15-43b7-bd2f-91370fcfefd6 | Starting:  Enter context: `users`" What does it mean? can you explain?
09:59 openstackgerrit Oleh Anufriiev proposed stackforge/rally: Publishing functional tests html reports  https://review.openstack.org/147098
10:03 rvasilets I understand this
10:50 Miouge joined #openstack-rally
11:06 yingjun joined #openstack-rally
11:24 oanufriev1 joined #openstack-rally
11:35 prmtl joined #openstack-rally
12:01 msdubov_ joined #openstack-rally
12:25 openstackgerrit Yuriy Nesenenko proposed stackforge/rally: Added list_snapshots scenario for Cinder test  https://review.openstack.org/146580
12:35 cdent joined #openstack-rally
12:47 openstackgerrit Mikhail Dubov proposed stackforge/rally: [WIP] Restructure Rally docs  https://review.openstack.org/120018
12:50 yingjun joined #openstack-rally
12:56 boris-42 joined #openstack-rally
12:58 pboros joined #openstack-rally
13:03 openstackgerrit Arx Cruz proposed stackforge/rally: Add a validator for neutron networks  https://review.openstack.org/141813
13:05 arxcruz gate-rally-dsvm-neutron-rally is failing a lot :/
13:09 psd joined #openstack-rally
13:29 psd_ joined #openstack-rally
13:33 mwagner_lap joined #openstack-rally
13:35 openstackgerrit Arx Cruz proposed stackforge/rally: Add a validator for neutron networks  https://review.openstack.org/141813
13:42 arxcruz amaretskiy, thanks for the +1 :D
13:43 amaretskiy arxcruz good work
13:43 arxcruz your comments enlighten me a lot :)
13:43 amaretskiy :)
13:51 openstackgerrit Oleh Anufriiev proposed stackforge/rally: Publishing functional tests html reports  https://review.openstack.org/147098
14:01 openstackgerrit Yuriy Nesenenko proposed stackforge/rally: Added list_snapshots scenario for Cinder test  https://review.openstack.org/146580
14:02 prmtl joined #openstack-rally
14:08 openstackgerrit Mikhail Dubov proposed stackforge/rally: [WIP] Restructure Rally docs  https://review.openstack.org/120018
14:09 openstackgerrit Mikhail Dubov proposed stackforge/rally: [WIP] Restructure Rally docs  https://review.openstack.org/120018
14:20 nmagnezi joined #openstack-rally
14:20 openstackgerrit Arx Cruz proposed stackforge/rally: Adding dict.items() concatenate validation  https://review.openstack.org/145458
14:21 dpaterson joined #openstack-rally
14:46 openstackgerrit Arx Cruz proposed stackforge/rally: Adding FunctionalMixin class  https://review.openstack.org/146838
14:59 openstackgerrit Yuriy Nesenenko proposed stackforge/rally: Added list_snapshots scenario for Cinder test  https://review.openstack.org/146580
15:07 jjmb joined #openstack-rally
15:07 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the generic script-based benchmark for the VMs  https://review.openstack.org/141671
15:07 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the context benchmark_image  https://review.openstack.org/138466
15:07 jjmb joined #openstack-rally
15:17 tnurlygayanov_ hi there
15:18 tnurlygayanov_ looks like 'rally deployment check' doesn't work on my environment
15:18 tnurlygayanov_ this command returns error:
15:18 tnurlygayanov_ Authentication Issues: wrong keystone credentials specified in your endpoint properties. (HTTP 401).
15:20 boris-42 arxcruz: hi
15:20 arxcruz boris-42, hello
15:20 boris-42 arxcruz: I left comments on Patch set 1
15:21 boris-42 arxcruz: there are some issues
15:21 arxcruz boris-42, which patch ?
15:21 arxcruz boris-42, oh, okay
15:22 boris-42 arxcruz: functional
15:22 arxcruz boris-42, gotcha, I'll make the changes :)
15:22 boris-42 tnurlygayanov_: what did you did?
15:23 boris-42 tnurlygayanov_: now if you would like to run Rally against MOS you need to speicfy "endpoint" argument
15:23 boris-42 arxcruz: as well create a Dummy benchmark
15:23 boris-42 arxcruz: that will check that this stuff works
15:24 tnurlygayanov_ boris-42 sorry, my bad, tenant name was incorrect
15:28 tnurlygayanov_ I specified endpoint='public' but I got error: Unable to establish connection to http://192.168.0.4:35357/v2.0/OS-KSADM/services  - why Rally tried to get admin endpoint?
15:29 boris-42 tnurlygayanov_: because it's fucking MOS
15:29 boris-42 =)
15:30 tnurlygayanov_ ok.... (
15:31 tnurlygayanov_ we need to fix it at least in Rally :) or it is required to change endpoints?
15:34 openstackgerrit Yuriy Nesenenko proposed stackforge/rally: Added list_snapshots scenario for Cinder test  https://review.openstack.org/146580
15:34 boris-42 еtnurlygayanov_ https://review.openstack.org/#/c/146490/4
15:34 boris-42 tnurlygayanov_: MOS is a bit thorugh ass
15:34 boris-42 tnurlygayanov_: so it's actually not a bug of Rally imho
15:35 boris-42 tnurlygayanov_: but that patch allows you to use rally and set managemnt url explicity
15:38 Miouge joined #openstack-rally
15:39 openstackgerrit Alexander Maretskiy proposed stackforge/rally: Optional parameter "endpoint" is added to deployment configuration  https://review.openstack.org/146490
15:44 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add new script for gate jobs  https://review.openstack.org/144633
15:55 tnurlygayanov_ boris-42 Thank you!
16:00 jjmb joined #openstack-rally
16:17 jjmb joined #openstack-rally
16:27 tnurlygayanov_ new problem...
16:28 tnurlygayanov_ mamy errors like: 2015-01-14 16:21:02.306 7054 WARNING rally.common.broker [-] Failed to consume a task from the queue: Gateway Timeout (HTTP 504)
16:28 tnurlygayanov_ after the benchmark
16:31 akuznetsova joined #openstack-rally
16:31 yfried joined #openstack-rally
16:31 rook__ joined #openstack-rally
16:34 igormarnat joined #openstack-rally
16:37 boris-42 tnurlygayanov_: please
16:37 boris-42 tnurlygayanov_: put whole exceptions
16:38 boris-42 tnurlygayanov_: not just a Messages from them
16:38 boris-42 tnurlygayanov_: + I belive you killed your Keysotne
16:38 boris-42 tnurlygayanov_: it's not rally problem
16:41 openstackgerrit Oleh Anufriiev proposed stackforge/rally: Publishing functional tests html reports  https://review.openstack.org/147098
16:41 akuznetsova joined #openstack-rally
16:44 igormarnat joined #openstack-rally
16:59 rook__ joined #openstack-rally
17:02 exploreshaifali joined #openstack-rally
17:16 jlk boris-42: are you around? I want a brain to bounce something off of
17:17 jlk boris-42: I'm running a boot-list test, constant. Times is 50, concurrency is 7
17:17 jlk My cloud has enough core capacity for 36 VCPUs, and nothing else is using it
17:17 jlk toward the end of the test, I'm getting a lot of no valid host errors, and scheduler log is showing 0 hosts returned from CoreFilter
17:19 jlk boris-42: even though rally is only doing 7 concurrent, I wonder if because it's so fast, that something is going wrong with the host checking in with what resources it has? and that number is getting inflated somehow?
17:20 amaretskiy jlk please give the full error from nova scheduler logs - scheduler must save an error message with reason of failure (for example, not enough memory,etc)
17:20 jlk /var/log/nova/nova-scheduler.log:2015-01-14 16:59:28.274 26048 INFO nova.filters [req-afdf8361-dd70-49f4-95ed-94eb7a553569 None] Filter CoreFilter returned 0 hosts
17:24 jlk yeah if I move the times to be lower than my capacity it works.
17:29 boris-42 jlk: hm
17:29 boris-42 jlk: but it will create 50 VMs
17:29 boris-42 jlk: so 36 VCPUs < 50
17:29 jlk even with a concurrency of 7?
17:30 boris-42 jlk: concurrency doesn't matter
17:30 jlk maybe I forgot some of this, it's been a while
17:30 boris-42 jlk: you are just booting VM
17:30 boris-42 jlk: you don't delete them
17:30 leeantho joined #openstack-rally
17:30 jlk ok, I need a boot - delete test :)
17:30 jlk so they're deleted as soon as they're booted
17:30 boris-42 jlk: it is already implemented
17:30 jlk nod
17:30 boris-42 jlk: boot-and-list is used to benchmark list() on scale
17:30 boris-42 jlk: like how it is changed when you have 1 vm 10 vm 100vms 1000vms=)
17:30 jlk yeah I copied the wrong test. My mistake
17:31 boris-42 jlk: btw try someday with fake driver=)
17:31 meteorfox joined #openstack-rally
17:32 boris-42 jlk: you'll see something like this http://logs.openstack.org/53/146953/1/check/gate-rally-dsvm-glance/5168aa2/rally-plot/results.html.gz#/GlancePlugin.create_and_list
17:32 boris-42 meteorfox: hey hey hey
17:32 jlk nod
17:33 boris-42 meteorfox: haven't see you for ages!
17:33 meteorfox boris-42: hey! yeah I know
17:33 boris-42 meteorfox: how are you?
17:34 meteorfox boris-42: turns out, I want to use Rally, for some performance tests
17:34 meteorfox boris-42: doing well, what about you?
17:34 boris-42 meteorfox: I am good as well=) a bit overbusy=)
17:34 boris-42 but oterwise good+)
17:34 meteorfox boris-42: cool
17:34 boris-42 meteorfox: btw we finished cleanup stuff
17:35 boris-42 meteorfox: so now they are plugable, with build-in-repeat mechanism and so on
17:35 boris-42 and they are well paralyzed
17:35 meteorfox boris-42: hehe.. cool.
17:35 boris-42 very important stuff for production envs=)
17:36 meteorfox boris-42:  yup, did you mean parallelized, as in parallel?
17:36 boris-42 meteorfox: yep I reinvent broker mechaism=)
17:36 boris-42 meteorfox: in rally
17:36 boris-42 to make it very optimal
17:36 meteorfox cool
17:37 boris-42 meteorfox: so this is plugin for example https://github.com/stackforge/rally/blob/master/rally/benchmark/context/cleanup/resources.py#L61-L63
17:37 meteorfox boris-42: right
17:37 boris-42 meteorfox: and it cleanups servers
17:38 boris-42 meteorfox: if you need somethign very custom
17:38 boris-42 meteorfox: you can take a look at neutron port https://github.com/stackforge/rally/blob/master/rally/benchmark/context/cleanup/resources.py#L113
17:38 boris-42 meteorfox: and here is new mechanism https://github.com/stackforge/rally/blob/master/rally/benchmark/context/cleanup/manager.py
17:39 meteorfox boris-42: I still don't get how the whole plugin thing works, but looks useful. I'll take a look later
17:39 boris-42 meteorfox: so this classes are autodiscovered
17:40 meteorfox boris-42: so, I have a "usage" question  about Rally
17:40 boris-42 meteorfox: sure
17:41 meteorfox boris-42: currently I have a private environment I want to test, and I have admin tenant. The thing is that Rally be default takes the adminURL from the service catalog
17:41 meteorfox boris-42: can I force rally to use, for example, the publicURL instead?
17:41 boris-42 meteorfox: publicURL won't allow you to run actions like creating usres and tenants
17:42 boris-42 meteorfox: that are required for rally
17:42 boris-42 meteorfox: but maybe this https://review.openstack.org/#/c/146490/
17:42 boris-42 meteorfox: can help you
17:42 meteorfox boris-42: i see
17:44 meteorfox so, endpoint overwrites adminURL?
17:44 boris-42 meteorfox: managment network
17:44 boris-42 meteorfox: yep
17:44 boris-42 meteorfox: so keystone won't generate it and will use that directly
17:45 meteorfox boris-42: I see, sounds like this is what I need
17:45 jlk huh
17:45 jlk even with boot-delete, and a concurrency of 20, I'm seeing up to 25~27 VCPUs used
17:49 meteorfox boris-42: what's endpoint_type used for?
17:52 tfreger joined #openstack-rally
17:53 meteorfox boris-42: is that changeset ready for merge?
18:00 boris-42 jlk: lag of quotas is about 1 minute
18:00 boris-42 jlk: resources manager*
18:00 boris-42 jlk:  so that can happen
18:00 boris-42 jlk: it's not rally issue=)
18:01 jlk nod
18:01 boris-42 meteorfox:  that one yep
18:01 boris-42 meteorfox: it's in process of merging
18:01 boris-42 meteorfox: I think endpointType will be removed at all in future..
18:03 meteorfox boris-42: so, let me see if I understood. With endpoint, I can tell rally to use that URL to create the tenants and the users?
18:03 meteorfox instead of using the adminURL
18:03 jlk that's how it works
18:03 boris-42 meteorfox: somekindofthat
18:03 jlk endpoint will override for "management url"
18:03 jlk it's a keystone object thing
18:03 jlk keystone will use the "management url" (or endpoint if it's passed along) for admin level actions
18:04 jlk otherwise it'll use what it gets out of the service catalog
18:05 meteorfox I see, so it authenticates with the auth_url, and for admin level actions uses endpoint url, if passed along, right?
18:05 boris-42 meteorfox: yep
18:06 meteorfox boris-42: cool. Then I think it will work for me too. Because I can auth just fine, and talk to compute just fine, my problem is just when Rally needs to create the tenants and users.
18:12 jlk so I have to question
18:12 jlk why don't you have the management url set right in the service catalog ?
18:13 meteorfox it is set, the thing it has an IPv4 I cannot access from my network, but the publicURL has the right one
18:14 meteorfox jlk: ^^
18:14 jlk then how do you access it?
18:15 meteorfox jlk: we have proxy, that only proxies for the publicURL
18:15 jlk why not set the proxy as the management url?
18:17 meteorfox I don't have the access to set it. They told me this is being set somewhere in a F5 LB conf :(
18:20 meteorfox jlk: here's what identity returns back for keystone, in my env, http://paste.openstack.org/show/xhI3Fw5aAXicW0tn6PGH/
18:22 jlk okay, if you can't change it, you can't change it, and the endpoint set will work for you
18:22 meteorfox jlk: right. thanks for helping me
18:31 meteorfox boris-42: also thanks for your help. it seems this change set is what I need. I'll wait for it to be merged, I hope is really soon.
18:36 boris-42 meteorfox|afk: you can cherry pick it
19:00 rook__ joined #openstack-rally
19:09 plieb joined #openstack-rally
19:12 nmagnezi joined #openstack-rally
19:54 dpaterson joined #openstack-rally
19:56 mwagner_lap joined #openstack-rally
20:00 exploreshaifali joined #openstack-rally
20:00 rook__ joined #openstack-rally
20:17 openstackgerrit Boris Pavlovic proposed stackforge/rally: Input task templates and task cmd cleanup  https://review.openstack.org/137716
20:32 meteorfox joined #openstack-rally
20:56 meteorfox boris-42: are you there?
20:56 boris-42 meteorfox: yep
20:57 meteorfox boris-42: so I tried the changeset, it's overriding the management url, but it's also using it as the auth_url
20:57 meteorfox is that expected
20:57 meteorfox ?
20:58 boris-42 meteorfox: not sure=)
20:58 boris-42 meteorfox: not sure that this will work
20:58 boris-42 meteorfox: you can try..
21:03 pboros_ joined #openstack-rally
21:16 plieb joined #openstack-rally
21:26 rook joined #openstack-rally
21:39 exploreshaifali joined #openstack-rally
21:59 msdubov_ joined #openstack-rally
22:35 openstackgerrit Oleh Anufriiev proposed stackforge/rally: WIP Publishing functional tests html reports  https://review.openstack.org/147098
23:15 rook joined #openstack-rally
23:16 vkmc boris-42, o/
23:19 boris-42 vkmc: hi hi
23:19 boris-42 vkmc: how are you?
23:19 vkmc boris-42, good and you? :)
23:19 boris-42 vkmc: good good=)
23:20 boris-42 vkmc: facing some races in openstack+)
23:20 boris-42 vkmc: trying to unblock rally gates lol
23:21 vkmc boris-42, boooo
23:21 vkmc boris-42, link?
23:21 vkmc boris-42, I dunno why I get so anxious when there is a gate broken
23:21 vkmc haha
23:22 boris-42 vkmc: it's not about gates
23:22 boris-42 vkmc: it's about races in various project
23:22 boris-42 vkmc: like heat & neutron broke today gate
23:22 boris-42 vkmc: rally gate*
23:22 boris-42 vkmc: and both neutron and heat are passing
23:22 boris-42 so I am saying wtf??
23:23 vkmc boris-42, oh thats odd
23:23 vkmc daaaamn
23:23 vkmc maybe there is a dependency updatE?
23:23 boris-42 vkmc: nope
23:23 boris-42 vkmc: it's just a code of projects
23:24 boris-42 vkmc: I hope this will fix everything https://review.openstack.org/#/c/147328/
23:50 openstackgerrit Oleh Anufriiev proposed stackforge/rally: WIP Publishing functional tests html reports  https://review.openstack.org/147098
23:52 vkmc boris-42, w00t, it looks so innocent
23:53 vkmc I take my hat off for the one that debugged that heh
23:53 boris-42 vkmc: =)
23:53 boris-42 vkmc: btw
23:53 boris-42 vkmc: what about trove benchmarks for rally?
23:54 vkmc boris-42, actually, that is a good question
23:54 vkmc boris-42, hope to put my hands on that soon... I have been busy playing with dib

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