Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2015-01-07

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

All times shown according to UTC.

Time Nick Message
00:05 barra204 joined #openstack-rally
00:35 yaguang joined #openstack-rally
00:48 rook__ joined #openstack-rally
01:17 yaguang joined #openstack-rally
02:04 rook__ joined #openstack-rally
03:53 rook__ joined #openstack-rally
04:01 mwagner_lap joined #openstack-rally
04:01 openstackgerrit Prabhjyot Singh Sodhi proposed stackforge/rally: Adding documentation  https://review.openstack.org/144835
04:02 psd joined #openstack-rally
04:50 psd joined #openstack-rally
05:24 rdas joined #openstack-rally
05:33 prmtl joined #openstack-rally
05:42 rook_ joined #openstack-rally
05:48 neeti joined #openstack-rally
06:02 nkhare joined #openstack-rally
06:02 tfreger joined #openstack-rally
06:20 rdas joined #openstack-rally
06:40 prmtl_ joined #openstack-rally
07:10 rook_ joined #openstack-rally
07:27 prmtl joined #openstack-rally
08:03 arxcruz joined #openstack-rally
08:59 rook_ joined #openstack-rally
09:03 yingjun joined #openstack-rally
09:12 pcaruana joined #openstack-rally
09:16 openstackgerrit Wataru Takase proposed stackforge/rally: Add list_volumes scenario for Cinder test  https://review.openstack.org/142500
09:17 pboldin joined #openstack-rally
09:30 prmtl joined #openstack-rally
09:35 openstackgerrit Arx Cruz proposed stackforge/rally: Python 3 fix str split For some reason, the str.split('\n') in python 3 is returning a space in the end of the line, while in python 2 doesn't. To fix it, I'm calling str.rstrip() to remove the trailing space. This will fix tests.unit.benchmark.processing  https://review.openstack.org/145444
09:35 arxcruz duh :/
09:35 openstackgerrit Arx Cruz proposed stackforge/rally: Python 3 fix str split  https://review.openstack.org/145444
09:35 arxcruz better :)
09:39 stannie joined #openstack-rally
09:46 openstackgerrit Arx Cruz proposed stackforge/rally: Python 3 fix '+' operand in dict in test_osclients.py  https://review.openstack.org/145450
10:09 andreykurilin joined #openstack-rally
10:12 yingjun joined #openstack-rally
10:13 yingjun joined #openstack-rally
10:27 openstackgerrit Arx Cruz proposed stackforge/rally: Adding dict.items() concatenate validation  https://review.openstack.org/145458
10:35 andreykurilin arxcruz: hi!
10:35 arxcruz andreykurilin, hello
10:36 andreykurilin arxcruz: how are you?
10:36 arxcruz great, and you ?
10:37 andreykurilin not bad, thanks:)
10:37 arxcruz :)
10:37 nkhare joined #openstack-rally
10:38 andreykurilin arxcruz: Do you have a bit free time?) we need one patch, which should unblock our gate-rally-python27
10:38 arxcruz andreykurilin, sure
10:39 andreykurilin arxcruz: please, look at https://github.com/stackforge/rally/blob/master/tests/unit/cmd/test_cliutils.py#L157-L160
10:40 andreykurilin arxcruz: this decorator should be moved from testcase method to whole class
10:41 andreykurilin arxcruz: until new oslo.config will be released, we don't have another way to fix our gates:(
10:48 rook_ joined #openstack-rally
10:52 andreykurilin joined #openstack-rally
10:54 openstackgerrit Arx Cruz proposed stackforge/rally: Adding dict.items() concatenate validation  https://review.openstack.org/145458
10:55 arxcruz andreykurilin, okay, I'll check that :)
11:06 boris-42 joined #openstack-rally
11:17 prmtl joined #openstack-rally
11:18 openstackgerrit Arx Cruz proposed stackforge/rally: Skipping CliUtilsTestCase  https://review.openstack.org/145470
11:18 arxcruz andreykurilin1, ^
11:39 arxcruz andreykurilin1, boris-42 http://paste.openstack.org/show/155834/
11:39 arxcruz If i change self.res_is.mock.assert_has_calls(mock.call('active')) to self.res_is.mock.assert_has_calls([mock.call('active')]) tests pass in both python 2 and 3
11:40 arxcruz but I'm not sure if that's the properly way to do that, or just a workaround
11:46 openstackgerrit Merged stackforge/rally: Python 3 fix '+' operand in dict in test_osclients.py  https://review.openstack.org/145450
11:47 cdent joined #openstack-rally
12:25 openstackgerrit Wataru Takase proposed stackforge/rally: Add list_volumes scenario for Cinder test  https://review.openstack.org/142500
12:30 openstackgerrit Arx Cruz proposed stackforge/rally: Adding dict.items() concatenate validation  https://review.openstack.org/145458
12:37 rook_ joined #openstack-rally
12:39 boris-42 arxcruz: hi
12:39 boris-42 arxcruz: could you put more changes in patches related for py34
12:39 boris-42 arxcruz: like 20-50 LOC per patch..
12:40 boris-42 arxcruz: and thank you for working on py34 compability
12:41 arxcruz boris-42, okay, I was trying to fix one problem per patch, I didn't saw the os_client error, otherwise I would put in the same patch, only saw today :)
12:42 boris-42 arxcruz: you can fix few (if they are small)
12:42 openstackgerrit Arx Cruz proposed stackforge/rally: Skipping CliUtilsTestCase  https://review.openstack.org/145470
12:42 boris-42 arxcruz: we should think about users that like to use "git log" to understand what happen in project
12:43 arxcruz boris-42, okay
12:43 arxcruz I'll keep that in mind :)
12:44 boris-42 arxcruz: thanks
12:44 arxcruz boris-42, regarding this http://paste.openstack.org/show/155834/ what do you think?
12:47 boris-42 arxcruz: seems like wrong mocking =)
12:48 arxcruz boris-42, okay, fixing :)
12:55 yingjun joined #openstack-rally
13:05 pboros joined #openstack-rally
13:07 openstackgerrit Arx Cruz proposed stackforge/rally: Python 3 fix mocking  https://review.openstack.org/145492
13:15 rook Morning!
13:16 arxcruz ahoj :)
13:25 boris-42 rook: HEY there
13:25 boris-42 rook: happy new year and so on
13:27 andreykurilin joined #openstack-rally
13:33 rook boris-42: Happy New year!
13:33 rook crazy hiatus..
13:37 rook_ joined #openstack-rally
13:56 boris-42 rook: =))
13:57 rook hows the lifting going?
14:14 mwagner_lap joined #openstack-rally
14:20 openstackgerrit Merged stackforge/rally: Skipping CliUtilsTestCase  https://review.openstack.org/145470
14:22 boris-42 rook: good -(
14:22 boris-42 =)
14:23 boris-42 rook: bench 75x10x4, dead lift 105x4x3, squat 65x10x4)
14:23 boris-42 rook: so will try to do deadlift 105x4x4
14:23 boris-42 rook: and squat 70x8x4 this week=)
14:24 rook Nice!!
14:24 psd_ joined #openstack-rally
14:24 rook I have been working on my gymnastics
14:24 rook 30 Muscle ups for time, 4min 58sec
14:25 boris-42 rook: ah that is as well very important
14:25 boris-42 rook: I'll be back in 2-3 hrs=)
14:25 boris-42 rook: see you later=)
14:26 rook aight, take care!
14:35 exploreshaifali joined #openstack-rally
14:44 prmtl_ joined #openstack-rally
15:02 nkhare joined #openstack-rally
15:08 rook_ joined #openstack-rally
15:40 neeti joined #openstack-rally
15:41 neeti joined #openstack-rally
15:45 psd_ joined #openstack-rally
16:05 exploreshaifali joined #openstack-rally
16:18 flwang joined #openstack-rally
16:38 openstackgerrit Wataru Takase proposed stackforge/rally: Add name pattern filter for resource cleanup  https://review.openstack.org/139643
16:39 rook__ joined #openstack-rally
17:06 rook__ joined #openstack-rally
17:30 psd boris-42, Hey!
17:35 boris-42 psd: what's up?
17:36 psd Nothing much, just had dinner!
17:37 * vkmc lurks
17:39 psd boris-42, okay regarding the adding duration of task... I think I will need to change print_list function in cliutils.py in rally/openstack/common
17:40 psd boris-42, print_list is used in task.py, but currently this function has been written so as to use just the current attributes like created_at and updated_at
17:41 psd so If I want to introduce duration, I can do it by making some changes in cliutils.py . But this seems to be more of a workaroud!
17:41 psd do I make sense in my doubt?
18:16 boris-42 Psd add new atteibute...
18:17 boris-42 Psd and display it
18:17 boris-42 So you won't need to change anything..
18:39 psd boris-42, Oh okay, can you please elaborate more upon how should I go about it
18:48 vkmc psd, is always a good idea to propose a change that you think is the best
18:48 vkmc reviewers will give you suggestions
18:53 psd vkmc, I believe what boris-42 is suggesting will be most easy to implement and will also maintain uniformity in code, but I am not sure how should I go about it, as in how to add an attribute to the task object. I tried searching in the code. I think I will have to read up sqlalchemy for understanding it :)
18:53 psd vkmc, but thanks
18:55 vkmc psd, boris-42 probably suggested you the way its done for other attributes... based on that, try to understand how its done checking out the code
18:56 openstackgerrit Jesse Keating proposed stackforge/rally: Don't hard set keystone endpoint  https://review.openstack.org/144366
19:00 psd vkmc, That's what! I am not able to find their code.
19:03 psd vkmc,  I trace it back to a IMPL.<some_fucntion call> where IMPL = db_api.DBAPI.from_config(CONF ...)
19:17 mwagner_lap joined #openstack-rally
19:17 pss joined #openstack-rally
19:19 rook__ joined #openstack-rally
19:24 pss joined #openstack-rally
19:31 boris-42 psd: sorry was offline
19:31 boris-42 psd: you have thing that you are putting to print_list
19:31 boris-42 psd: update it to add duration
19:32 boris-42 psd: and put to print_list
19:32 boris-42 psd: it's really the simplest way and the way that we need
19:36 psd boris-42, I can add duration to the headers (no problem doing this) but when it fetches value from the database, I don't get how to add the new attribute 'duration'
19:45 jlk pip timeouts to openstacks' mirrors.
19:45 jlk sigh
19:50 andreykurilin joined #openstack-rally
20:01 psd|away joined #openstack-rally
20:08 boris-42 jlk: lol
20:08 boris-42 jlk: it happens
20:09 boris-42 jlk: btw could you upadte your patch one more time..
20:10 boris-42 psd modife values
20:10 boris-42 psd|away: that you got from DB
20:10 boris-42 psd|away:  https://github.com/stackforge/rally/blob/master/rally/cmd/commands/task.py#L371
20:10 boris-42 psd|away: add to every element of list new field with duration
20:16 boris-42 jlk: btw
20:16 boris-42 jlk: you Operator of OpenStack?
20:17 boris-42 jlk: could you review a proposal for adding profiler for OpenStack: https://review.openstack.org/#/c/134839/
20:17 boris-42 jlk: it need some love
20:20 rook__ joined #openstack-rally
20:31 tnurlygayanov_ joined #openstack-rally
20:32 jlk will take a look
20:44 exploreshaifali joined #openstack-rally
20:45 jlk oh. yeah. That battle. I'm not sure I want to wade into that, particularly since I have history with Sandy
21:06 openstackgerrit joined #openstack-rally
21:31 boris-42 jlk: ahahaha
21:32 boris-42 jlk: history with Sandy olol
21:32 boris-42 =)
21:32 jlk we used to work together at RAX :)
21:32 boris-42 jlk: sorry about that LOL
21:32 boris-42 )
21:33 jlk he's a nice enough guy on the whole
21:36 rook__ joined #openstack-rally
21:43 frayedknot joined #openstack-rally
21:47 boris-42 jlk: yep I know
21:47 boris-42 jlk: but sometimes..
21:47 boris-42 jlk: when he talks about stacktach..
21:47 jlk yeah
21:47 jlk it's his baby
21:48 boris-42 jlk: yep yep
21:49 boris-42 frayedknot: hi there=)
21:49 boris-42 jlk: btw https://review.openstack.org/#/c/145595/ this should help cinder team
21:49 boris-42 jlk: to fix one of the races that we are caching in rally gates
21:50 frayedknot hello
21:50 boris-42 frayedknot: are you working on cray?
21:58 boris-42 rook: okay I am here
21:59 boris-42 rook: btw you have one patch in rally that requires a bit of love
21:59 boris-42 rook: could you address comment https://review.openstack.org/#/c/115439/ ?
22:08 rook boris-42: sure, i will get to it - in all day meetings tomorrow.
22:09 rook might need to look at the linpack code anyway
22:09 boris-42 rook: meetings booo
22:09 boris-42 =)
22:09 boris-42 jlk: some magic doesn't allow us to merge your patch
22:10 jlk hahah ahahaha, whoo.
22:10 jlk should I resubmit?
22:10 rook boris-42 i need to add some vmtask work I have done that is a bit more comprehensive when networks are involved.
22:11 boris-42 jlk: nono
22:11 boris-42 jlk: so actually mostly it is because some critical bug return to Cinder
22:11 boris-42 jlk: and strikes again
22:11 boris-42 so we need a bit of luck to avoid it in one of run..
22:11 boris-42 rook: btw
22:12 boris-42 rook: paboldin continued work on tzbal patches
22:12 boris-42 rook: and seems like he want to finish it=)
22:13 rook awesome - i need to sync again with Marcio re, vmtask + cloudbench. I have been swamped with new baby and work
22:13 rook ;/
22:16 jlk boris-42: oh, okay.
22:16 boris-42 rook: btwI forgot to ask about baby
22:16 boris-42 rook: she he??
22:16 boris-42 rook: yep integration of cloudbanch seems like a nice thing
22:18 rook he :D
22:18 rook http://www.talerico-photography.com/Family/Baby-Coles-Pictures/
22:18 rook have to go feed the wife, catch you later boris-42
22:19 boris-42 rook: oh nice nice
22:19 boris-42 rook: see you
22:35 jlk WELP
22:36 jlk Havana is definitely vulnerable to the glance bug.
22:36 boris-42 jlk: lol
22:36 boris-42 jlk: I don't know how people is using OS in production)
22:37 boris-42 jlk: I can DDOS any cloud using keystone.authenticate benchmark
22:37 boris-42 jlk: from my phone
22:37 jlk oh sure.
22:37 boris-42 jlk: http://logs.openstack.org/85/136485/3/check/gate-rally-dsvm-keystone/93012ee/rally-plot/results.html.gz
22:37 jlk rate limiters get in the way of that though
22:37 boris-42 jlk: take a look at graphs ^ =)
22:38 boris-42 jlk: it is just about 2k of authentication
22:38 jlk RAX puts in some pretty heavy rate limiters
22:38 boris-42 jlk:  I can create N user s
22:38 boris-42 jlk: and DDoS it=)
22:38 jlk no, because the rate is tied to the whole tenant
22:38 jlk you'd have to have N number of individual Rackspace public cloud accounts
22:38 boris-42 jlk: but various users - various tenants?
22:38 jlk no
22:38 boris-42 jlk: so why I can't have N public cloud accounts?)
22:38 jlk each rackspace public cloud account is it's own tenant, and users exist within that tenant
22:39 jlk you could
22:39 boris-42 jlk: so I can use them and ddos cloud?)
22:39 jlk you would have to have N valid credit card numbers and go through N valid on-boarding
22:39 boris-42 jlk: ya they called me after registration
22:39 jlk adn even then, the rate limiters are pretty heavy, I think you'd have to have a lot of tenants
22:39 boris-42 jlk: and asked my why I did that lol
22:39 boris-42 =)
22:40 jlk for us, each customer gets their own whole cloud, so the worst they could do is disrupt their own cloud
22:40 jlk which... awesome, good job :)
22:40 boris-42 jlk: =)
22:40 boris-42 jlk:  I am not hacker=) but I could try to do that lol=)
22:40 boris-42 jlk: is that legal?)
22:41 jlk legal?  Gray area.
22:41 boris-42 =)
22:41 boris-42 jlk: so take a look at that graph!=)
22:41 jlk nobody thus far has been motivated enough to attack RAX public cloud in that way
22:41 jlk then again, RAX public cloud doesn't use keystone.
22:41 boris-42 jlk: hm
22:41 jlk it implements the keystone API
22:41 boris-42 jlk: but what about internal services?
22:41 jlk through ohter software
22:42 boris-42 jlk: ah that is smart=)
22:42 jlk heh
22:42 jlk well... it's a giant pile of proprietary java code....
22:47 boris-42 jlk: heh
22:48 boris-42 jlk: btw did you mention some bugs related to security groups
22:48 boris-42 jlk: like you are creating VM adding sec groups
22:48 boris-42 jlk: and they are not added
22:48 boris-42 jlk: cause we are facing this in gates as well
22:48 boris-42 and on our scale lab in Mirantis
22:49 jlk no I don't think I've seen that recently
22:49 jlk s/recently/
22:49 jlk /
22:49 boris-42 jlk: strange maybe it's new feature=)
22:50 boris-42 jlk: this stuff is failing https://github.com/stackforge/rally/blob/master/rally-jobs/rally.yaml#L1119-L1141
22:51 jlk ah, I don't think I'm running that here yet
23:25 rook__ joined #openstack-rally
23:33 nkhare joined #openstack-rally
23:54 boris-42 jlk: okay seems like whole gates are down
23:55 boris-42 nice
23:55 boris-42 jlk: specially for your patch lol

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