Perl 6 - the future is here, just unevenly distributed

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

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

All times shown according to UTC.

Time Nick Message
00:02 marcoemorais1 joined #openstack-rally
00:03 marcoemorais1 joined #openstack-rally
00:15 marcoemorais joined #openstack-rally
00:36 marcoemorais joined #openstack-rally
00:36 marcoemorais joined #openstack-rally
00:37 marcoemorais joined #openstack-rally
00:50 marcoemorais joined #openstack-rally
00:58 yingjun joined #openstack-rally
01:00 yingjun joined #openstack-rally
01:06 yingjun joined #openstack-rally
01:19 yaguang joined #openstack-rally
01:38 mwagner_lap joined #openstack-rally
01:44 boris-42 joined #openstack-rally
01:58 jjmb joined #openstack-rally
02:00 mwagner_lap joined #openstack-rally
02:23 Viswanath joined #openstack-rally
02:31 yingjun joined #openstack-rally
02:36 jjmb joined #openstack-rally
02:45 jjmb joined #openstack-rally
02:50 erkules joined #openstack-rally
02:59 jjmb joined #openstack-rally
03:00 yingjun joined #openstack-rally
03:30 chandan_kumar joined #openstack-rally
03:42 debu_ joined #openstack-rally
03:47 klindgren_ joined #openstack-rally
04:20 mcaulfie joined #openstack-rally
04:35 AlexF joined #openstack-rally
04:36 mcaulfie joined #openstack-rally
05:04 rvcehimanshu joined #openstack-rally
05:22 rdas joined #openstack-rally
05:22 neeti joined #openstack-rally
05:36 marcoemorais joined #openstack-rally
05:38 marcoemorais1 joined #openstack-rally
05:39 k4n0 joined #openstack-rally
06:36 yingjun joined #openstack-rally
06:45 openstackgerrit Mikhail Dubov proposed a change to stackforge/rally: Refactor UserGenerator  https://review.openstack.org/119297
06:45 chandankumar joined #openstack-rally
06:45 chandankumar joined #openstack-rally
06:51 rdas joined #openstack-rally
07:29 svasheka joined #openstack-rally
08:23 openstackgerrit Timur Nurlygayanov proposed a change to stackforge/rally: Adding Network Context Class  https://review.openstack.org/103306
08:23 jjmb joined #openstack-rally
08:30 AlexF joined #openstack-rally
08:34 openstackgerrit Mikhail Dubov proposed a change to stackforge/rally: Refactor UserGenerator  https://review.openstack.org/119297
08:49 nkhare joined #openstack-rally
09:03 amaretskiy joined #openstack-rally
09:07 oanufriev joined #openstack-rally
09:08 tnurlygayanov Hi there
09:08 tnurlygayanov I have fixed my commit https://review.openstack.org/#/c/103306/ but I can see fails of tests in gate-rally-dsvm-rally
09:09 tnurlygayanov and looks like the root of this problem not in commit, but in race condition in gate job
09:18 ehaselwanter joined #openstack-rally
09:19 amaretskiy tnurlygayanov hi
09:19 tnurlygayanov amaretskiy hi :)
09:19 amaretskiy take a look at http://logs.openstack.org/06/103306/71/check/gate-rally-dsvm-rally/767a80e/console.html
09:19 amaretskiy 2014-11-11 08:52:49.055 | 2014-11-11 08:52:49.021 | 2014-11-11 08:52:48.956 25621 TRACE rally TypeError: _create_network() got an unexpected keyword argument 'project_id'
09:20 amaretskiy it seems that we hav ereal issue
09:20 amaretskiy and we catch it on our gates
09:24 amaretskiy tnurlygayanov also I see ZeroDivisionError
09:31 tnurlygayanov amaretskiy thank you! looks like the root of this fail in my code :)
09:32 AlexF joined #openstack-rally
10:01 fandi joined #openstack-rally
10:11 boris-42 joined #openstack-rally
10:31 openstackgerrit Timur Nurlygayanov proposed a change to stackforge/rally: Adding Network Context Class  https://review.openstack.org/103306
11:40 Harikrishna joined #openstack-rally
11:46 rdas joined #openstack-rally
11:47 openstackgerrit Alexander Maretskiy proposed a change to stackforge/rally: Base HTML templates and improvements for task report  https://review.openstack.org/131844
11:55 andreykurilin joined #openstack-rally
12:02 sponomarenko joined #openstack-rally
12:11 cdent joined #openstack-rally
12:25 nkhare joined #openstack-rally
12:27 openstackgerrit Timur Nurlygayanov proposed a change to stackforge/rally: Adding Network Context Class  https://review.openstack.org/103306
12:29 pboros joined #openstack-rally
12:34 pboros joined #openstack-rally
12:51 yingjun joined #openstack-rally
12:56 andreykurilin joined #openstack-rally
12:58 nkhare joined #openstack-rally
13:02 andreykurilin yingjun, hi!
13:12 mwagner_lap joined #openstack-rally
13:14 yingjun andreykurilin hi
13:20 rediskin joined #openstack-rally
13:21 rvcehimanshu joined #openstack-rally
13:26 andreykurilin yingjun: how are you?
13:27 yingjun andreykurilin, good, thanks
13:45 andreykurilin yingjun: I have one more task related to oslo code. Do you interested in it?
13:48 yingjun andreykurilin, if it isn’t urgent
13:50 andreykurilin yingjun: yeah, It is not urgent, but important. Can you sync the latest code from oslo-incubator to Rally? Do you know how to do it?
13:50 yingjun andreykurilin, yes
13:51 andreykurilin yingjun, good:)
13:51 yingjun andreykurilin, i can do it maybe in couple of days
13:51 andreykurilin yingjun, great! thank you
14:01 boris-42 yingjun:  hey hey
14:01 boris-42 yingjun:  did you have time to visit Paris?)
14:05 yingjun boris-42, actually i did walk around pairs, it’s really a nice place
14:05 yingjun boris-42, still in paris?
14:06 boris-42 yingjun: nope I just back today
14:07 yingjun boris-42, good
14:07 boris-42 yingjun:  me ready to work now=)
14:08 yingjun boris-42, yeah, me too
14:11 openstackgerrit Timur Nurlygayanov proposed a change to stackforge/rally: Adding Network Context Class  https://review.openstack.org/103306
14:21 mcaulfie joined #openstack-rally
14:37 fandi joined #openstack-rally
14:42 openstackgerrit Sergey Murashov proposed a change to stackforge/rally: Add benchmarks for murano  https://review.openstack.org/127491
14:55 boris-42 ekarlso: hey there
15:15 paboldin joined #openstack-rally
15:19 marcoemorais joined #openstack-rally
15:20 rvcehimanshu joined #openstack-rally
15:23 marcoemorais1 joined #openstack-rally
15:36 jjmb joined #openstack-rally
16:04 jjmb joined #openstack-rally
16:27 klindgren joined #openstack-rally
16:30 paboldin hi guys. I want to enhance Rally with functionality that would aid us in semi-automatic performance tuning.
16:30 paboldin What is the process for the changes? Should I start with a blueprint?
16:32 amaretskiy paboldin hi
16:33 boris-42 paboldin: we like new contributors
16:33 amaretskiy paboldin you create a blueprint, but that is not mandatory
16:33 boris-42 paboldin: and we don't need blueprints actually
16:33 boris-42 paboldin:  but if you wish you can do it, only thing is that we need to discuss what actually are you going to do
16:34 boris-42 paboldin:  to avoid any issues
16:34 paboldin boris-42: ok, so I need to write a draft first.
16:36 boris-42 paboldin:  https://review.openstack.org/#/c/125238/23/rally/utils.py heh
16:36 paboldin boris-42: sorry about that
16:36 boris-42 paboldin:  ya or just share use case idea
16:36 rediskin left #openstack-rally
16:36 boris-42 paboldin:  no good catch
16:36 boris-42 paboldin:  thanks=)
16:36 boris-42 paboldin:  could you make patch that fix that lol?)
16:36 boris-42 paboldin:  if you find such stuff in code just propose patches if you have time=)
16:36 paboldin boris-42: just to go through the process? ok
16:36 KimJ joined #openstack-rally
16:37 boris-42 paboldin:  no we care a lot about quality of code
16:37 paboldin boris-42: i see. it is fine-developed. Most of the code is decorators, lol :-)
16:37 boris-42 paboldin:  lol=)
16:37 KimJ joined #openstack-rally
16:37 boris-42 paboldin: it will be even more lol
16:37 boris-42 paboldin:  =)
16:38 boris-42 paboldin:  but please if you find any nits in code that you dislike =)
16:38 boris-42 paboldin:  just publsih patches
16:38 boris-42 paboldin:  code can be always better and we are using approach of continious refactoring
16:38 paboldin boris-42: ok
16:39 boris-42 paboldin:  as well it will be interesting to see proposal of what you want to implement
16:39 boris-42 paboldin: at least just to understand use cases
16:39 paboldin boris-42: i only know what I want it to do, not how it can be done.
16:39 paboldin yeah, for now only use cases
16:39 boris-42 paboldin:  that is why we have feature request
16:40 boris-42 paboldin:  you can propose https://github.com/stackforge/rally/tree/master/doc/feature_request here
16:40 boris-42 paboldin: only thing that you need is to write short description of use cases that you would like to cover
16:40 paboldin boris-42: ok
16:40 boris-42 paboldin:  then we will start discussing how to get it done
16:40 paboldin boris-42: great!
16:41 boris-42 paboldin:  ya it's a bit simpler than specs=)
16:41 boris-42 paboldin:  you don't need to push any tech details=)
16:41 dmellado hey boris-42, thanks for having a chat with me at the summit
16:41 boris-42 dmellado: np =)
16:41 dmellado did the guy from enovance contacted you about the network testing functionality?
16:41 boris-42 dmellado:  edeploy?
16:42 dmellado iirc he was erwan_taf at freendoe
16:42 dmellado freenode
16:42 boris-42 dmellado: yep
16:42 boris-42 dmellado:  I didn't have enough time to investigate everything
16:42 boris-42 dmellado:  but I started already
16:42 boris-42 dmellado:  btw we will make a google doc
16:42 boris-42 dmellado:  where we will start huge discussion about what to do
16:42 dmellado great, please do so so we can discuss it
16:42 boris-42 paboldin:  so I'll wait for your patch
16:43 paboldin boris-42: ok
16:43 boris-42 dmellado:  yep and we will share it via openstack-dev and I'll ping you
16:43 chandankumar joined #openstack-rally
16:44 dmellado great, ty!
16:44 mcaulfie joined #openstack-rally
16:50 AlexF joined #openstack-rally
16:51 vhosakot joined #openstack-rally
16:51 vhosakot hi, for a nova scenario, how do I see the output of "self._list_servers(detailed=True)" ?
16:52 vhosakot I'm looking for the output of "nova list".. I don't see it in stsout
16:52 vhosakot should I enable rally debug -d option to see the outputs ?
16:53 Viswanath joined #openstack-rally
16:55 msdubov joined #openstack-rally
16:58 amaretskiy vhosakot hi
16:58 amaretskiy vhosakot there is no saved output of `nova list'
16:59 amaretskiy because that is not a purpose of scenario
16:59 oanufriev joined #openstack-rally
16:59 amaretskiy the purpose is to measure execution of `nova list'
17:00 vhosakot amaretskiy: thanks for your reply... I agree there is no purpose from a nova scenario perspective.... how do I see the networks associated with the VMs booted ?
17:00 ekarlso boris-42: ello
17:00 amaretskiy you can use nova CLI
17:01 amaretskiy vhosakot you can use openrc created by rally for that
17:01 amaretskiy example:  $ . ~/.rally/openrc ; nova net list
17:02 amaretskiy something like that
17:05 dmellado in fact he could use it his keystonerc, isn't it? rally gets a copy of them
17:10 vhosakot amaretskiy: I know about using the CLI... but, I wanted to know which networks were the servers booted with for the nova scenario case.. is there a way to find this ?
17:13 marcoemorais joined #openstack-rally
17:13 amaretskiy vhosakot now there is no auto-spawned networks in rally master, but there is a patch which is on its way to master and it implements networking context both for nova and neutron - https://review.openstack.org/#/c/103306/
17:14 amaretskiy if you want to play with this patch - note that networks created by rally will be deleted after scenario is complete
17:14 vhosakot amaretskiy: I learned in this chatroom yesterday that all the nova boot scenarios use one of the pre-created networks in the output of "rally show networks".. is there a way to find which of the pre-created networks does rally use for a nova boot scenario ?
17:14 amaretskiy i think you can get list these networks by nova cli during the scenario run time, if you want
17:15 amaretskiy vhosakot you must specify this network exactly in the scenario file :)
17:16 vhosakot amaretskiy: oh really.. where is the scenario file, please :)
17:18 vhosakot amaretskiy: where is the scenario file ?
17:19 amaretskiy wait please
17:20 amaretskiy sample from rally/doc/... missed such parameter
17:20 amaretskiy https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/nova/utils.py#L91
17:22 amaretskiy also https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/nova/utils.py#L90
17:22 andreykurilin joined #openstack-rally
17:23 jjmb joined #openstack-rally
17:23 vhosakot amaretskiy: thanks, I will refer them...
17:23 amaretskiy vhosakot you ban specify network id in scenario args: kwargs['nics'] = [{'net-id': net.id}]
17:23 jjmb joined #openstack-rally
17:24 amaretskiy s/ban/can/
17:25 vhosakot amaretskiy: cool, can I also mention in boot-and-delete.json too under "args" ?
17:26 vhosakot amaretskiy: I should also have auto_assign_nic set to true right ?
17:27 amaretskiy no, if you have auto_assign_nic = True, then rally will find proper network by itself
17:28 amaretskiy you can just specify exact network in scenario args
17:28 amaretskiy network id
17:28 vhosakot amaretskiy: oh ok ok.. I will have auto_assign_nic = false, and pass the the net id in scenario args then...
17:28 vhosakot amaretskiy: cool, I will try it, and let you know.. thanks!
17:28 amaretskiy uto_assign_nic is false by default
17:29 amaretskiy just specify net-id
17:30 amaretskiy JSON: ... "args": { "nics": [{"net-id": "here-is-network-uuid"}]            ... }
17:30 vhosakot amaretskiy: wow, cool.. thanks a lot!
17:30 amaretskiy np :)
17:31 vhosakot amaretskiy : net-id will be passed onto _boot_server thru kwargs right ?
17:31 amaretskiy yes
17:32 amaretskiy that is missed in samples because we don't know exact net-id so sample will require modification for run :)
17:38 leeantho joined #openstack-rally
17:44 vhosakot amaretskiy : I passed the net-id in the json, and see an error when I run...... Error <class 'novaclient.exceptions.BadRequest'>: Network fe1edfa7-1dae-44a2-a645-e7991989add9 could not be found. (HTTP 400)
17:45 vhosakot amaretskiy : the network fe1edfa7-1dae-44a2-a645-e7991989add9 does exist, and seen in "rally show networks"
17:47 amaretskiy vhosakot make net-show for this network and paste it to http://paste.openstack.org/
17:48 openstackgerrit joined #openstack-rally
17:50 amaretskiy vhosakot also you can run rally --debug ... and paste part of the output as well
17:52 vhosakot amaretskiy :    http://paste.openstack.org/show/131970/  has the output of networks, json, and command output
17:54 vhosakot amaretskiy: boot-and-delete.json has under "args".....   "nics": [{"net-id": "fe1edfa7-1dae-44a2-a645-e7991989add9"}]
17:54 boris-42 vhosakot: hi
17:55 vhosakot boris-42: hi, how was your flight, and how is laptop battery ?
17:55 amaretskiy vhosakot try public network id
17:56 vhosakot boris-42: was it you who said yesterday that you had to run to catch a flight and your laptop battery is 1% charge :)
17:56 vhosakot amaretskiy: let me public net-id and let you know... 2 mins please
17:58 vhosakot amaretskiy: I see that the option for nova boot is nic.... is nics (plural) correct with a list of multiple net-id's ?
17:58 boris-42 vhosakot:  now it's much better conncted to charger=)
17:58 oanufriev joined #openstack-rally
17:59 vhosakot amaretskiy: I see that the option for nova boot is nic.... is nics (plural) correct with a list of multiple net-id's ?
17:59 vhosakot boris-42: cool, glad you connected the charge and your power is back :)
17:59 amaretskiy yes, that is correct
17:59 vhosakot amaretskiy: cool, cool :)
18:00 vhosakot amaretskiy: Now, I see this error for public net-id........    Error <class 'rally.exceptions.GetResourceErrorStatus'>: Resource <Server: rally_novaserver_cbcqhgpptddzjyjr> has ERROR status: No valid host was found.
18:00 amaretskiy vhosakot now try smaller flavor :)
18:01 boris-42 vhosakot: this is message from nova scheduler
18:01 boris-42 vhosakot: run "nova list --all-tenants"
18:01 boris-42 vhosakot:  probably there are some active VMs that were not deleted or something like that
18:01 amaretskiy No valid host was found usually means we do not have enough openstack resources to start VM
18:01 amaretskiy yes
18:01 vhosakot "nova list --all-tenants" is empty and has not output
18:02 boris-42 vhosakot: so then probably you need to try smaller tenant..
18:02 boris-42 flavor**
18:02 boris-42 sry
18:02 boris-42 pboros: hi there
18:03 fandi joined #openstack-rally
18:04 pboros hi boris-42
18:04 vhosakot amaretskiy:   I am using m1.nano in my json... which is smaller than nano... micro, tiny ?
18:04 vhosakot amaretskiy: I see many flavors in "rally show flavors", but cannot decide which is smaller than m1.nano
18:05 boris-42 vhosakot:  hm hm
18:06 boris-42 pboros: how are you? still in Paris?)
18:06 vhosakot amaretskiy: it looks like m1.nano is the smallet
18:06 vhosakot smallest*
18:06 boris-42 vhosakot: seems like something wired is with your cloud
18:06 boris-42 vhosakot: can you just try to boot vm from CLI?
18:07 vhosakot boris-42: good idea, let me try nova boot and pass the nic
18:07 boris-42 vhosakot:  ya cause for some reason nova scheduler can't find proper host
18:08 pboros no, I am at home already
18:08 pboros how are you?
18:12 oanufriev joined #openstack-rally
18:12 vhosakot amaretskiy, boris-42: "nova boot" CLI works fine, and I am able to boot a server...
18:12 vhosakot nova boot --flavor m1.nano --image cirros-0.3.2-x86_64-uec --nic net-id=16788445-07b6-444a-9ae8-dd7afeee37fd test_server.... worked fine
18:13 boris-42 vhosakot: hmhmhmhm
18:13 boris-42 vhosakot:  could you share rally input task
18:14 boris-42 pboros: ah just came back thing morning=)
18:14 msdubov joined #openstack-rally
18:16 boris-42 vhosakot:  that fails
18:16 vhosakot boris-42: what exactly is the input task ? json file ?
18:16 boris-42 vhosakot:  so the one reason that can produce such behaviour is next
18:16 boris-42 vhosakot:  rally creates VMs from "temp" created tenants
18:17 boris-42 vhosakot:  that are not admin, and if network is not shared it wont' work
18:18 vhosakot boris-42: oh ok... so, rally does not create as admin ?
18:18 boris-42 vhosakot:  nope
18:18 boris-42 vhosakot:  even if you don't pass "users" context it will pass it by default
18:19 boris-42 vhosakot:  it's for now, when we finish LDAP, it will print validation errors in such errors
18:19 boris-42 vhosakot:  actually in rally we are trying to drop "admin" requiremts.
18:19 boris-42 vhosakot: for most of benchmarks, that actually don't use it.. But it is huge task=)
18:19 boris-42 vhosakot:  and we are working for a long period time on it..
18:22 vhosakot boris-42:  I have pasted my json and the existing.json I passed to "rally deployment create" at   http://fpaste.org/149765/73004314/.... Could you tell me if there is anything wrong / inconsistent ?
18:22 boris-42 pboros: so are you ready to rock?)
18:23 boris-42 vhosakot:  the thing with network
18:23 pboros boris-42, yes:)
18:23 boris-42 vhosakot:  are you sure that it's not amdin only
18:23 boris-42 vhosakot:  and you can actually use it?
18:24 vhosakot boris-42: I have username as "admin" in existing.json
18:24 boris-42 vhosakot:  one more time
18:25 boris-42 vhosakot:  it wont' use it to create VM
18:25 boris-42 vhosakot:  it will create temp tenants & users and will use them to generate load
18:25 boris-42 pboros: do you need help with some of steps?)
18:25 boris-42 pboros: actually this may help you https://wiki.openstack.org/wiki/Rally/Develop#How_to_contribute
18:26 vhosakot boris-42: everything works fine when I do not pass the nics parameter in "args" in the json
18:26 pboros boris-42, I probably will, this week I am a bit post-conference swamped, but next week I already allocated some time for rally
18:27 boris-42 vhosakot:  you are specifing "admin only nic"
18:27 boris-42 vhosakot:  I belive
18:27 boris-42 vhosakot: that's why it fails
18:27 boris-42 vhosakot:  if you create some shared network you'll be able to use it. Actually that's why we are working on NetworkContext
18:27 boris-42 pboros: great
18:29 vhosakot boris-42: so, I need to share the network associated with the nic (passed to _boot_server) with the the temp user rally creates to run the scenarios... if yes, what is the temp username rally creates and uses I need to share the network with ?
18:32 boris-42 vhosakot: every time differnt
18:33 boris-42 vhosakot:  that is why we need to finish https://review.openstack.org/#/c/103306/
18:33 vhosakot boris-42: wow, so what do I do then... :)  how do I associate a server booted by rally to a specific nic
18:36 boris-42 vhosakot: so what I thougth is that if network is shared
18:36 boris-42 vhosakot:  you can use it from any tenant?
18:37 boris-42 vhosakot:  so the answer is to make your network "shared"
18:37 vhosakot boris-42: cool, let me make the network shared... need to search the neutron CLI for it... thanks, will do it, and let you know... :)
18:38 boris-42 vhosakot:  sure, I am as well not big specialist in neutron=)
18:40 vhosakot boris-42: cool, first.. I will go running very fast, grab some lunch, and come back running very fast, and continue working...  will let you know once I make the network shared.. I think I need to pass the shared parameter in the json to neutron net-update
18:41 boris-42 vhosakot:  sounds good =)
18:47 boris-42 ekarlso: are you around?
18:48 boris-42 ekarlso:  can you fix designate benchmarks in rally, you made a change in designate client
18:48 boris-42 ekarlso: using keystone authentication session or something like that
18:48 boris-42 ekarlso:  that brakes everything
18:49 jjmb joined #openstack-rally
18:58 marcoemorais joined #openstack-rally
18:59 marcoemorais joined #openstack-rally
19:06 openstackgerrit Pavel Boldin proposed a change to stackforge/rally: Removing redundancy in (a==b) and 1 or 0  https://review.openstack.org/133807
19:06 paboldin boris-42: sorry, was busy testing + setting up py26 tests
19:07 boris-42 paboldin: hehe=)
19:07 boris-42 paboldin:  sometimes it produces issues=)
19:07 paboldin boris-42: yeah, but i have a centos in the chroot
19:10 ekarlso boris-42: sorry ? that shouldn't affect anything ?
19:11 boris-42 ekarlso: ya but it affected
19:12 ekarlso boris-42: example ?
19:12 boris-42 I just recently run in rally gates check experimental
19:12 boris-42 this is output http://logs.openstack.org/60/129060/31/experimental/gate-rally-dsvm-designate-rally/a57ad9e/console.html
19:13 paboldin boris-42: why no tests are being run at the moment?
19:13 boris-42 paboldin:  hm?
19:14 paboldin at my request. or they don't start automagically?
19:14 boris-42 paboldin:  they start
19:14 boris-42 paboldin:  they work for 30 minutes
19:14 boris-42 ~ 30 mins*
19:14 paboldin oh, and there is no 'started' message in the gerrit?
19:15 boris-42 paboldin:  you can track it here if you would like http://status.openstack.org/zuul/
19:17 boris-42 paboldin: a far as I remember there were
19:17 boris-42 paboldin:  but at some point they dissapared
19:17 paboldin yeah, there are
19:17 boris-42 actually they are useless
19:19 paboldin https://jenkins01.openstack.org/job/gate-rally-dsvm-neutron-rally/266/
19:19 paboldin why?
19:19 paboldin there is only unit-tests?
19:22 boris-42 paboldin: hm?
19:22 paboldin boris-42: there is no tests of rally tasks start ?
19:22 marcoemorais joined #openstack-rally
19:22 boris-42 paboldin:  why do you think so?)
19:22 paboldin boris-42: i'm just asking :)
19:23 boris-42 paboldin: so nope
19:23 boris-42 paboldin:  in case of only unit tests rally won't be so stable=)
19:23 boris-42 paboldin:  before juno we didn't have functional gates =)
19:23 boris-42 paboldin:  so you can belive me=)
19:24 paboldin ok
19:24 boris-42 gate-rally-install* jobs tests rally/install.sh script
19:25 boris-42 gate-rally-dsvm-=rally and gate-rally-dsvm-neutron tests actually task start
19:25 vhosakot boris-42: I'm back from lunch... made the network shared... the neutron CLI is....   neutron net-update <network_name> --shared
19:25 boris-42 and a bit other commnds (actually all task related commnds)
19:25 boris-42 paboldin: gate-rally-dsvm-cli covers most of commands that you have in CLI
19:25 paboldin yeah, already got that from the name :)
19:26 boris-42 paboldin: gate-rally-tox-self will be isolated job for rally testing
19:26 boris-42 paboldin:  on scale* and perfromance
19:26 boris-42 like just running dummy scenarios 1kk times with 1k concurency
19:26 boris-42 and checking that nothing is broken*
19:27 boris-42 running 10k times rally task start and then doing list and so on
19:27 boris-42 killing rally in the middle and testing graceful shutdown
19:27 boris-42 and so on=)
19:27 boris-42 paboldin:  so we are testing a lot =)
19:28 boris-42 paboldin: and going to test much more=)
19:28 paboldin boris-42: cheers for coverage :)
19:28 boris-42 paboldin:  heh not enough we need 146%!
19:28 paboldin :D
19:30 ehaselwanter joined #openstack-rally
19:31 vhosakot boris-42: are you there :)
19:31 boris-42 vhosakot: ya
19:32 boris-42 vhosakot:  lemme gues it doens't work?)
19:32 vhosakot boris-42: noooooo.. your guess is wrong... allow me to type :)
19:32 boris-42 vhosakot: =)))
19:32 vhosakot boris-42: yes, it worked after I shared the network... no http error this time. and the scenario worked fine
19:33 vhosakot boris-42: :)
19:33 boris-42 vhosakot:  awsome
19:33 boris-42 vhosakot:  actually if want to help rally
19:33 vhosakot boris-42: how do I see the VMs list that were booted in th test
19:33 boris-42 vhosakot:  this is very old bug
19:33 boris-42 vhosakot:  and you can fix it
19:33 boris-42 vhosakot:  heh we don't store such info
19:34 boris-42 vhosakot: my cat is eating my usb cable lol!
19:34 vhosakot boris-42: in boot_and_delete_server, if I comment out self._delete_server part, I should see the servers created in "nova list" output after the test ends,,,
19:34 boris-42 vhosakot: nope=)
19:35 vhosakot boris-42: hahaha, I used to feed whisky for my dog
19:35 boris-42 vhosakot: sooo
19:35 boris-42 vhosakot: there is generic cleanup stuff in rally
19:35 vhosakot boris-42: why, if the servers are not deleted, won't they remain after the test
19:35 boris-42 vhosakot:  nope they will be cleaned by cleanup context
19:35 boris-42 vhosakot: btw there is already benchmark that just boots vms=)
19:36 vhosakot boris-42: then, I need to add code myself between boot server and delete server to see server list... unfortunately self._list_servers does not print anything
19:36 boris-42 vhosakot: ya cause it shouldn't..
19:36 vhosakot boris-42: (I will answer about my contribution to rally in a bit)
19:36 boris-42 vhosakot:  so I mean why do you need that info?
19:37 boris-42 vhosakot:  I mean you can add LOG.debug() or something else
19:37 boris-42 vhosakot: but the intersting for me question is why?)
19:37 vhosakot boris-42: but, between boot server and delete server, I can add code myself to print the servers right using NOVA APIs... right, using LOG.debug or the more general print statement
19:37 boris-42 don't use print * use log
19:37 boris-42 vhosakot:  yep you can do that
19:37 boris-42 vhosakot:  and it will work
19:38 vhosakot boris-42: or another faster hack.. is to add a LARGE sleep (like 10 mins or so) between and boot server and delete server, and ssh thru another window and issue NOVA CLIs :)
19:39 boris-42 vhosakot:  it is already there...
19:39 boris-42 vhosakot:  take a look here https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/nova/servers.py#L72
19:40 boris-42 vhosakot:  but as I said you can just LOG.debug
19:40 boris-42 vhosakot:  or patch self._list_servers()
19:40 vhosakot boris-42: cool... self.sleep_between(min_sleep, max_sleep)  will do the work for me...
19:40 boris-42 vhosakot:  https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/nova/utils.py#L71-L74
19:40 boris-42 vhosakot: if you need to exececute something on VM
19:40 boris-42 vhosakot:  there is benchmark for that..
19:41 boris-42 vhosakot:  not sure what you are trying to do =) so not able to help=)
19:42 vhosakot boris-42: basically... I want to use rally to boot some VMs associated with a network that has DHCP enabled, and and measure the time needed by each VM to obtain the DHCP IP address...
19:42 boris-42 vhosakot: so you actually can use benchmark for that
19:42 boris-42 vhosakot:  or write own.. actually
19:42 vhosakot boris-42:  I have resolved this part now -- booting VMs associated with network that has DHCP enabled
19:43 boris-42 vhosakot:  so actually I was looking for somebody that will make new extra argument (wait_for_ip)
19:43 vhosakot boris-42: once the VMs are booted, I want to use NOVA CLIs to check what DHCP IP address they are assigned
19:43 boris-42 vhosakot:  in benchmarks related to NovaServers
19:43 boris-42 vhosakot:  why not do it in rally benchmark?
19:43 mcaulfie joined #openstack-rally
19:44 vhosakot boris-42: well, yes it can be made as a benchmark, in the long term yes
19:44 boris-42 vhosakot: believe it will be faster then what you are trying to do
19:44 vhosakot boris-42: so, now, I have 4 VMs booted associated with a network that has DHCP enalbed
19:44 jjmb joined #openstack-rally
19:44 boris-42 vhosakot: btw take a look at this https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/vm/vmtasks.py#L45
19:45 boris-42 especially this https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/vm/vmtasks.py#L86
19:45 boris-42 and this parts
19:45 boris-42 https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/vm/vmtasks.py#L91-L96
19:45 boris-42 it's almost what you are looking for
19:45 vhosakot boris-42: I don't need to login / ssh into each VM and issue "ifconfig" to check the DHCP IP assigned to it... NOVA CLIs will show what DHCO IP a VM has been issued
19:46 boris-42 vhosakot: so you missed fully my point
19:46 boris-42 vhosakot: take that code and use it in another benchmark
19:46 boris-42 vhosakot:  that just run vm, assing IP (if it is required), waits for ping
19:46 boris-42 vhosakot:  all code is already done in utils
19:47 boris-42 vhosakot:  you need just to make few lines of code that will call proper utils methods
19:47 boris-42 =)
19:47 vhosakot boris-42: let me check the code.. thanks.. will let you know... :)
19:47 boris-42 vhosakot:  so what I am looking for in long term
19:47 boris-42 vhosakot:  is to add extra argument to all NovaServers benchmarks
19:47 boris-42 vhosakot:  like wait_for_ping
19:48 boris-42 vhosakot:  that will do what you are trying to do actually..
19:48 vhosakot boris-42: yes,,, right
19:48 boris-42 vhosakot: so that can be already easily combinied from existing code=)
19:49 boris-42 vhosakot:  that's how rally evolves
19:49 vhosakot boris-42: :)
19:50 vhosakot boris-42: I would love to contribute to rally (only if I had the time)... will think about it...
19:51 vhosakot boris-42:  how can I send the args  min_sleep, max_sleep ?
19:51 boris-42 vhosakot: just put them in args section..
19:53 vhosakot boris-42: cool... trying it now
19:53 marcoemorais joined #openstack-rally
19:56 boris-42 paboldin: so
19:57 boris-42 paboldin: it passes gates=)
20:02 jjmb joined #openstack-rally
20:35 jjmb joined #openstack-rally
21:06 AlexF joined #openstack-rally
21:09 Viswanath joined #openstack-rally
21:44 vhosakot boris-42: how can I measure the CPU and memory usage for nova/neutron scenarios ? I remember we could measure CPU and memory for rally scenarios by passing them as args in the json file
21:45 boris-42 vhosakot: what? =)
21:45 boris-42 vhosakot: CPU and memory of what? rally process?
21:46 vhosakot boris-42: yeah, CPU and memory usage for a rally scenario..
21:46 boris-42 vhosakot:  Idk=)
21:46 boris-42 vhosakot:  we don't do that=)
21:46 boris-42 vhosakot:  it shouldn't use too much actually..
21:47 vhosakot boris-42: oh ok... so rally does not measure CPU memory, right ? I just wanted to check if rally measures CPU and memory for scenarios
21:47 boris-42 vhosakot: I mean I am not sure that I fully get what you are asking
21:47 boris-42 vhosakot:  why do care about how much CPU and MEM consumes rally scenario?
21:48 vhosakot boris-42: well, things like how much CPU and memory was used to say boot 100 VMs
21:49 boris-42 vhosakot:  but it's imho question to OpenStack project
21:49 boris-42 vhosakot: projects* not to rally
21:49 boris-42 vhosakot: how rally can now that if it uses only OpenStack API
21:49 boris-42 know*
21:50 vhosakot boris-42: right right, I understand... just wanted to confirm... rally is amazing to scale nova and neutron, was just thinking if we can masure CPU and memory as well when we scale stuff
21:53 marcoemorais joined #openstack-rally
21:53 boris-42 vhosakot:  if we can get this info via OpenStack API then we can
21:55 boris-42 vhosakot:  if not then we can think about extending OpenStack API to support that
21:55 boris-42 vhosakot: cause without special agents on all compute/volume/.. nodes it's impossible
21:56 vhosakot boris-42: something like "nova usage-list"
21:56 vhosakot that shows instance usage stats
21:58 boris-42 vhosakot: so you can collect that data and return via scenario output
21:59 boris-42 vhosakot: like in VMTask
21:59 vhosakot boris-42: yeah, at the end... I thought Rally already did that... actually it is more like a celiometer task
22:03 boris-42 vhosakot:  you can use rally if you would like
22:04 vhosakot boris-42: yes, I am using rally now to scam nova and neutron... it is great to scale servers and networks
22:04 vhosakot boris-42: :)
22:09 boris-42 vhosakot:  actually it's interesting stuff
22:09 boris-42 vhosakot:  to fetch usage
22:10 vhosakot boris-42: yes, scale testers would like to see CPU and memory usage... that is why I asked :)
22:13 ekarlso boris-42: so, designate, what's broke ?
22:21 boris-42 ekarlso: heh
22:21 boris-42 ekarlso: so it doesn't work at all
22:21 ekarlso boris-42: uuuh, meaning ? :D
22:22 boris-42 ekarlso:  just push any dummy patch
22:22 boris-42 ekarlso: and trigger designate job
22:27 boris-42 ekarlso: or run locally
22:27 boris-42 ekarlso: any benchmark
22:28 ekarlso boris-42: example of a failing job ?
22:28 ekarlso can't see it on designate project
22:29 boris-42 ekarlso:  ?
22:29 ekarlso boris-42: trying to find a log of it failing on review.o.o ?
22:29 boris-42 ekarlso: ?
22:30 boris-42 ekarlso: I send you logs...
22:30 ekarlso oook :P
22:30 boris-42 ekarlso: already few times..
22:30 ekarlso uhm, sorry but i've missed them :(
22:31 boris-42 ekarlso: http://logs.openstack.org/60/129060/31/experimental/gate-rally-dsvm-designate-rally/a57ad9e/console.html
22:32 boris-42 ekarlso: actually can you put here SLA checks https://github.com/stackforge/rally/blob/master/rally-scenarios/rally-designate.yaml
22:32 boris-42 ekarlso: ?
22:33 boris-42 ekarlso: like this https://github.com/stackforge/rally/blob/master/rally-scenarios/rally.yaml#L11-L12
22:33 boris-42 ekarlso: so when you run next time "check experimental" we will see that it doesn't work
22:34 ekarlso k
22:34 oanufriev joined #openstack-rally
22:35 ekarlso boris-42: oh I see the bug ith that one
22:35 boris-42 ekarlso: and it will be nice to fix bug=)
22:36 boris-42 ekarlso: so we will get +1/-1 Jenkins for designate + working benchmarks for designate=)
22:40 cat2jackson joined #openstack-rally
22:43 ekarlso https://review.openstack.org/#/c/133838/1 boris-42  should fix it
22:43 boris-42 ekarlso: so it's not rally bug?
22:44 ekarlso boris-42: nope :D
22:44 boris-42 ekarlso: lol
22:44 ekarlso designateclient bug that I introduced when adding sessions :/
22:44 boris-42 ekarlso: btw I will still prefer if you add patch to rally
22:44 boris-42 ekarlso: with SLA checks
22:44 ekarlso sure :)
22:44 ekarlso I will master :)
22:44 boris-42 ekarlso: I actually found this bug accedentally
22:45 boris-42 ekarlso: cause I was refactoring cleanup stuff=)
22:45 ekarlso :P
22:49 openstackgerrit Endre Karlson proposed a change to stackforge/rally: Add SLA to Designate scenarios  https://review.openstack.org/133839
22:50 boris-42 ekarlso: =)
22:53 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 3  https://review.openstack.org/129060
22:53 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 2  https://review.openstack.org/132437
22:53 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 1  https://review.openstack.org/132436
22:53 boris-42 ^ ekarlso
22:53 boris-42 ekarlso: now it's tasty+)
22:59 ekarlso omg :p
23:02 boris-42 ekarlso: it was like one patch
23:02 boris-42 ekarlso:  but I thought that something like +2500 -1000 is bad =)
23:04 dmellado_ joined #openstack-rally
23:07 vhosakot boris-42: I'm going home now.. will chat tomorrow here.. thanks a lot for all the help :)
23:08 boris-42 vhosakot:  np
23:09 boris-42 ekarlso: so you see now it's failing https://review.openstack.org/#/c/133839/ =)
23:17 jjmb joined #openstack-rally
23:29 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 3  https://review.openstack.org/129060
23:30 dmellado_ joined #openstack-rally
23:48 mwagner_lap joined #openstack-rally

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