Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2014-12-24

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

All times shown according to UTC.

Time Nick Message
00:12 yingjun joined #openstack-rally
00:20 pboldin joined #openstack-rally
00:26 openstackgerrit Li Yingjun proposed stackforge/rally: Move files to common lib(Part 1)  https://review.openstack.org/143376
00:29 mwagner_lap joined #openstack-rally
00:32 dmorita joined #openstack-rally
01:02 flwang1 joined #openstack-rally
01:12 jjmb joined #openstack-rally
01:13 flwang1 joined #openstack-rally
01:13 mitz joined #openstack-rally
01:22 openstackgerrit li,chen proposed stackforge/rally: Use six.string_types instead of basestring  https://review.openstack.org/143585
01:30 openstackgerrit li,chen proposed stackforge/rally: Use six.string_types instead of basestring  https://review.openstack.org/143585
01:32 openstackgerrit li,chen proposed stackforge/rally: Use six.string_types instead of basestring  https://review.openstack.org/143585
01:40 openstackgerrit Merged stackforge/rally: Move files to common lib(Part 1)  https://review.openstack.org/143376
01:54 yaguang joined #openstack-rally
01:58 chen andreykurilin keep getting failed for pep8 test. :(    http://paste.openstack.org/show/154268/
01:59 openstackgerrit li,chen proposed stackforge/rally: Use six.string_types instead of basestring  https://review.openstack.org/143585
02:01 openstackgerrit li,chen proposed stackforge/rally: Use six.string_types instead of basestring  https://review.openstack.org/143585
02:22 openstackgerrit Li Yingjun proposed stackforge/rally: Move files to common lib(Part 2)  https://review.openstack.org/143347
02:45 yingjun joined #openstack-rally
03:37 Poornima joined #openstack-rally
05:01 rdas joined #openstack-rally
05:18 nkhare joined #openstack-rally
05:46 boris-42 joined #openstack-rally
06:10 boris-42 chen: I replied on your comment
06:11 boris-42 chen: in patch*
06:11 boris-42 chen: you should improve your regexp
06:11 chen boris-42 Thanks!
06:11 * boris-42 is going to sleep now
06:32 nmagnezi_ joined #openstack-rally
06:40 chen boris-42 ...   :(    couldn't understand that!
06:42 yingjun joined #openstack-rally
06:49 k4n0 joined #openstack-rally
07:09 boris-42 chen: ?
07:15 yfried joined #openstack-rally
07:17 yingjun_ joined #openstack-rally
07:24 rvcehimanshu joined #openstack-rally
07:29 chen boris-42 never mind~~ I figured that out.
07:30 yingjun joined #openstack-rally
07:37 yingjun joined #openstack-rally
07:40 yfried boris-42: ping
07:43 boris-42 yfried: hi there
07:44 boris-42 chen: if you have any questions the best way to get answer is to add comments to your patch*
07:44 boris-42 chen: but you can as well ping in any time and I'll reply when I'll be online
07:45 yfried boris-42: sure
07:45 yfried :
07:45 chen boris-42   sure  ~
07:45 boris-42 yfried: so what you would like to contribute?
07:45 yfried boris-42: well, I'm looking for "low-hanging-fruits" to help
07:47 yfried boris-42: I saw on the wiki you have plans to work better with tempest, maybe I could help there?
07:47 boris-42 yfried: ah that is okay
07:47 boris-42 yfried: yeeep
07:47 yfried boris-42: I'm also pretty familiar with Neutron/networkin
07:47 boris-42 yfried: btw there is one thing that you can improve in our Nova benchmarks
07:47 boris-42 yfried: related to networking
07:47 boris-42 yfried: add "wait_for_ping"
07:47 boris-42 yfried: argument
07:48 yfried boris-42: Nova?
07:48 boris-42 yfried: that thing was asked multple times and it's quite simple
07:48 boris-42 yfried: so this set of benchmarks https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/nova/servers.py
07:48 boris-42 yfried: that boots VM
07:48 chen boris-42   Just a little stupid question, for re.compile(r"xxx"), why there is an "r" in () before "regexp" ?
07:49 chen boris-42 read ?
07:49 boris-42 chen: nope
07:49 yfried chen: http://stackoverflow.com/questions/2081640/what-exactly-do-u-and-r-string-flags-do-in-python-and-what-are-raw-string-l
07:50 yfried chen: does that help?
07:50 boris-42 chen: here is good http://stackoverflow.com/questions/2081640/what-exactly-do-u-and-r-string-flags-do-in-python-and-what-are-raw-string-l answer
07:50 boris-42 yfried: heh)
07:50 boris-42 yfried: thanks
07:50 chen boris-42  yfried  Thank you both.  :)
07:50 chen boris-42  yfried  very helpful !!
07:50 boris-42 chen: so basically it means don't touch string cause it is regexp
07:50 boris-42 =)
07:51 boris-42 yfried: sooo
07:51 boris-42 yfried: we already actually have atomic action for "wait for ping"
07:51 boris-42 yfried https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/vm/utils.py#L45-L51
07:51 boris-42 yfried: so it makes sense to move it from vm/utils -> nova/utils
07:52 boris-42 yfried: and add argument "wait_for_ping" to all nova scenarios
07:52 boris-42 yfried: and if it is "wait_for_ping=True" call this method, if it is False not
07:52 boris-42 yfried: so we will check that actually our VMs are not only spawned but as well pingable
07:53 boris-42 which means that network is set properly
07:53 boris-42 yfried: how about that?
07:53 yfried boris-42: that's might be noob question but this assumes a network is ready and reachable?
07:53 boris-42 yfried: yep
07:54 yfried boris-42: where can I read about the difference between "vm" and "nova" trees?
07:54 boris-42 yfried: only code..
07:55 boris-42 yfried: but there is not so much code there
07:55 boris-42 yfried: actually "nova" tree is related more to benchmarking API
07:55 boris-42 yfried: VM is related to benchmarking inside VM
07:55 boris-42 yfried: like running iperf and so on
07:55 boris-42 yfried: that will be implemented in future
07:57 neeti joined #openstack-rally
07:57 yfried boris-42: can both trees share the ping code, or do I "copy-paste" this?
07:58 boris-42 yfried: I belive that you should move (not copy code) to Nova.utils
07:58 boris-42 yfried: to this class https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/nova/utils.py#L72
07:59 boris-42 yfried: and inherit vm/utils.py from it
07:59 boris-42 btw we should align this with paboldin
08:00 yfried boris-42: ok, I see.
08:00 boris-42 yfried: cause he is working on big refactoring of that part
08:00 yfried boris-42: do you have a bug/bp I could link to?
08:00 openstackgerrit li,chen proposed stackforge/rally: Use six.string_types instead of basestring  https://review.openstack.org/143585
08:00 boris-42 yfried: so basically we don't creat bugs for such thing, cause they are not bugs=)
08:01 boris-42 yfried: and we don't create BP cause it's small features
08:01 yfried boris-42: ok. just making sure
08:01 boris-42 and we are trying to avoid bureaucracy
08:01 boris-42 yfried: if you would like you can create a bug and I'll mark it to whishlist
08:01 boris-42 yfried: but just good written code with unit tests is enough for me
08:02 boris-42 yfried: and when andreykurilin wake up we can ping him about work on verfication part of rally
08:02 boris-42 yfried: he is main maintainer of that part
08:02 yfried boris-42: I'll start working on that
08:03 boris-42 yfried: sure great
08:03 boris-42 yfried: it shouldn't be too complicaed
08:03 yfried boris-42: doesn't sound like that.
08:04 boris-42 yfried: and it is quite coool stuff
08:17 rvcehimanshu joined #openstack-rally
08:19 yfried boris-42: a little question about the structure and code sharing:
08:19 boris-42 yfried: sure
08:20 pboldin joined #openstack-rally
08:20 yfried boris-42: so vs_utils.run_command is using wait_for_ping
08:20 yfried boris-42: but CMScenario only inherits base.Scenario
08:21 yfried VMScenario
08:21 boris-42 yfried: yep you need to change that
08:21 yfried boris-42: I guess
08:21 boris-42 yfried: VmScenario should inherit NovaScenario
08:21 yfried boris-42: is that allowed? great
08:21 boris-42 yfried: yep that is what you should do=)
08:22 boris-42 yfried: and remove NovaScenario from VMTask
08:22 boris-42 yfried: cause it will be already inherit in VMScenario
08:22 yfried boris-42: ack
08:22 yfried boris-42: and I'm replaceing nova as the vm's super (not adding to it). right?
08:27 rvcehimanshu joined #openstack-rally
08:34 rvcehimanshu joined #openstack-rally
08:36 boris-42 yfried: VMScenario(nova.utils.NovaScenario) VmTask(vm.utils.VMScenario)
08:36 boris-42 yfried: like that
08:57 pboldin joined #openstack-rally
09:00 yfried boris-42: tnx
09:01 yfried boris-42: IIUC, you want "_boot_server" to return when server is pingable instead of "ACTIVE" status, right?
09:07 yfried boris-42: are nested atomic actions allowed?
09:07 boris-42 yfried: it should have argument
09:07 boris-42 yfried: nope
09:07 boris-42 yfried: this can be implemented in various ways
09:08 boris-42 yfried: we can use with AtomicAction() construction inside _boot_server
09:08 boris-42 yfried: or call directly in benchmark scenarios wait_for_ping
09:08 boris-42 if wait_for_ping(): self._wait_for_ping()
09:09 boris-42 both are OK
09:11 yfried boris-42: I'm not sure I understand the 1st op. do you have an example I can look at?
09:11 boris-42 yfried: yep I have just a moment
09:12 boris-42 yfried: here is implementation https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/base.py#L245-L299
09:12 boris-42 yfried: and sample is
09:13 boris-42 yfried: https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/keystone/utils.py#L48-L52
09:17 rdas joined #openstack-rally
09:21 yfried boris-42: I'm sorry, I'm having a bit of trouble here
09:21 yfried boris-42: _boot_server is an atomic action, right?
09:22 yfried boris-42: and we want it to have to option of returning after ping instead of after "active" right?
09:25 openstackgerrit Merged stackforge/rally: Move files to common lib(Part 2)  https://review.openstack.org/143347
09:35 rmk joined #openstack-rally
09:35 rmk joined #openstack-rally
09:36 boris-42 yfried: remove decorator
09:36 boris-42 yfried: add inside code AtomicAction
09:36 boris-42 yfried: and put inside with construction evertyhing except call of _with_for_ping
09:37 yfried boris-42: oh
09:37 yfried boris-42: so boot would be 2 atomic actions?
09:38 boris-42 yfried: some kind of that
09:40 yfried boris-42: also - no floating IP on boot server - how can we ping the VMs?
09:41 boris-42 yfried: do in the similiar way
09:41 boris-42 yfried: if no floating try to ping fixedIP
09:41 boris-42 yfried: floatingIP can be attached if you specify nics and so on as input args
09:43 yfried boris-42: I don't think so. AFAIK, floating ip can't be created and associated as part of "nova boot"
09:44 boris-42 yfried: we can extend that as well it's python=)
09:44 boris-42 yfried: https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/vm/vmtasks.py#L94-L97 it's simple=)
09:45 yfried boris-42: tnx
09:45 yfried boris-42: should FLIP creation be a different atomic, or part of ping?
09:45 boris-42 yfried: actually we can make some ultamegacool method
09:45 boris-42 yfried: called boot_vm
09:46 boris-42 yfried: that will do all that stuff and that can be reused by both VMTask and nova scenarios
09:46 yfried boris-42: I'm listening, but you don't have to sell me on "ultamegacool" :) I'm already here
09:46 boris-42 yfried: I would prefer different
09:46 boris-42 yfried: =)
09:46 boris-42 yfried: so sorry but I have to move to office
09:47 boris-42 yfried: I'll be back in few hours
09:47 yfried boris-42: ok. tnx.
09:47 boris-42 yfried: if you have questions ping andreykurilin or redixin =)
09:47 yfried boris-42: what time zone are you?
09:47 boris-42 yfried: Moscow
09:47 yfried boris-42: tnx
09:47 boris-42 yfried: so it's UTC +4
09:51 openstackgerrit Merged stackforge/rally: Use six.string_types instead of basestring  https://review.openstack.org/143585
10:05 yfried andreykurilin: redixin: I'm new to rally. how can I run it in with debugger?
10:14 tkammer joined #openstack-rally
10:20 amaretskiy joined #openstack-rally
10:25 yingjun joined #openstack-rally
10:28 yingjun joined #openstack-rally
10:38 openstackgerrit li,chen proposed stackforge/rally: Use six.moves.StringIO instead of StringIO.StringIO  https://review.openstack.org/143584
10:42 openstackgerrit Mikhail Dubov proposed stackforge/rally: Restructure Rally docs (WIP)  https://review.openstack.org/120018
10:51 pboldin joined #openstack-rally
10:58 redixin yfried, you can run rally with debugger like any other python project
10:59 redixin yfried, or you want to run unit tests with debugger?
11:01 yfried redixin: I want to run the scenario with debugger
11:01 yfried redixin: in tempest I simply use nose
11:02 redixin yfried, you may add import pdb; pdb.set_trace() in the code
11:02 yfried redixin: and it will stop?
11:02 redixin it will stop script and start pdb
11:03 yfried redixin: are you using some IDE's debugger, or from console?
11:03 redixin yfried, im not using IDE, and almost never using pdb
11:03 redixin testing rules, debugging sucks =)
11:19 openstackgerrit Mikhail Dubov proposed stackforge/rally: Restructure Rally docs (WIP)  https://review.openstack.org/120018
11:27 pboldin joined #openstack-rally
11:27 openstackgerrit Mikhail Dubov proposed stackforge/rally: Restructure Rally docs (WIP)  https://review.openstack.org/120018
11:28 andreykurilin redixin: hi
11:28 redixin andreykurilin, sup
11:29 andreykurilin redixin: import pdb;pdb.set_trace no always works :( for example, in scenarios, which are launched in multiproccess mode
11:29 redixin yfried ^
11:31 yfried andreykurilin: what if I sec concurrency to 1 (or 0)?
11:32 boris-42 andreykurilin: yep
11:32 boris-42 andreykurilin: it doesn't work
11:32 boris-42 andreykurilin: that's why we added serial runner
11:32 andreykurilin yes
11:32 andreykurilin serial runner should be selected
11:32 yfried boris-42: so serial runner would enable debug?
11:34 boris-42 yfried: yep it should
11:42 openstackgerrit Mikhail Dubov proposed stackforge/rally: Restructure Rally docs (WIP)  https://review.openstack.org/120018
11:43 msdubov_ joined #openstack-rally
11:44 k4n0 joined #openstack-rally
11:58 redixin joined #openstack-rally
12:12 openstackgerrit Olga Kopylova proposed stackforge/rally: Decorator to mark deprecated benchmark components  https://review.openstack.org/138489
12:23 openstackgerrit Merged stackforge/rally: Use six.moves.StringIO instead of StringIO.StringIO  https://review.openstack.org/143584
12:23 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add Dockerfile  https://review.openstack.org/132556
12:29 openstackgerrit li,chen proposed stackforge/rally: raise syntax change for py3 incompatible  https://review.openstack.org/143831
12:39 gema joined #openstack-rally
12:44 jaypipes joined #openstack-rally
12:46 chen andreykurilin  what shuld  "[result_(i) for i in 0, 1, 2]" changed to ????  I find nothing at http://docs.pythonsprints.com/python3_porting/py-porting.html
12:47 andreykurilin chen: put "0, 1, 2" to tupple
12:48 andreykurilin chen: something like [result_(i) for i in (0,1,2)]
12:48 andreykurilin checn: this should work in both python 27 and  python 34
12:48 chen andreykurilin   o
12:49 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the flavor context  https://review.openstack.org/141670
12:49 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the generic script-based benchmark for the VMs  https://review.openstack.org/141671
12:49 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the context benchmark_image  https://review.openstack.org/138466
12:49 openstackgerrit Pavel Boldin proposed stackforge/rally: Tests for rally gates  https://review.openstack.org/143469
12:49 chen andreykurilin python3 do not support "," at everywhere ?
12:50 andreykurilin chen: probably yes
12:50 andreykurilin chen: python 3 is more strict:)
13:00 jjmb joined #openstack-rally
13:01 jjmb1 joined #openstack-rally
13:10 openstackgerrit Mikhail Dubov proposed stackforge/rally: [WIP] Restructure Rally docs  https://review.openstack.org/120018
13:11 openstackgerrit li,chen proposed stackforge/rally: syntax change for py3 incompatible  https://review.openstack.org/143831
13:12 oanufriev joined #openstack-rally
13:13 cdent joined #openstack-rally
13:22 vkmc joined #openstack-rally
13:22 vkmc joined #openstack-rally
13:35 openstackgerrit Mikhail Dubov proposed stackforge/rally: [WIP] Restructure Rally docs  https://review.openstack.org/120018
13:40 openstackgerrit Mikhail Dubov proposed stackforge/rally: [WIP] Restructure Rally docs  https://review.openstack.org/120018
13:42 yfried joined #openstack-rally
13:50 openstackgerrit Merged stackforge/rally: Add Dockerfile  https://review.openstack.org/132556
13:54 andreykurilin_ joined #openstack-rally
14:02 openstackgerrit Mikhail Dubov proposed stackforge/rally: [WIP] Restructure Rally docs  https://review.openstack.org/120018
14:09 jjmb joined #openstack-rally
14:10 jjmb joined #openstack-rally
14:10 jjmb2 joined #openstack-rally
14:11 boris-42 chen: nice work on py34!
14:27 openstackgerrit Merged stackforge/rally: syntax change for py3 incompatible  https://review.openstack.org/143831
14:37 openstackgerrit Yair Fried proposed stackforge/rally: Moves ping utilities from "vm" to "nova"  https://review.openstack.org/143849
14:38 yfried boris-42: ^ that's my work so far
14:38 yfried boris-42: I am trying to test it
14:38 yfried boris-42: but I think I'm missing something
14:39 boris-42 yfried: I will take a look soon
14:41 yfried boris-42: "rally task start" is using the files in "site-packages" and not from the local git repo, so I'm constantly having to run ./install_rally and recreate my deployment to test my work
14:43 yfried is there a scenario that creates a network for the VM or do I have to rely on a pre-created shared network?
14:46 openstackgerrit Boris Pavlovic proposed stackforge/rally: Refactor validation (part 1)  https://review.openstack.org/143743
14:50 boris-42 yfried: use Network Context
14:50 boris-42 yfried: amaretskiy finshed it recently
14:50 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano benchmarks  https://review.openstack.org/137661
14:50 boris-42 yfried: it creates neworks for every tenant
14:50 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano base for murano benchmarks  https://review.openstack.org/137650
14:50 yfried amaretskiy: boris-42: could you please point me to doc?
14:50 yfried tnx
14:51 boris-42 yfried: https://review.openstack.org/#/c/103306/
14:53 amaretskiy yfried we should refactor utils (both nova and neutron)
14:53 yfried amaretskiy: ?
14:53 amaretskiy and use our new network wrappers in utils
14:54 amaretskiy so we should move existent code or add new into rally.benchmark.wrappers.network
14:55 yfried amaretskiy: since I've just started looking at rally yesterday, I've no idea about new or existing code :)
14:55 amaretskiy and then use it in rally.benchmark.scenarios.(nova|neutron).utils
14:55 amaretskiy yfried i will just make a review of your patch set and put comments
14:55 amaretskiy so i hope everything will be clear :)
14:56 yfried amaretskiy: keep in mind that it's WIP
14:56 amaretskiy of course
14:56 amaretskiy let's step forward to merging to master, patch-set by patch-set :)
15:03 andreykurilin_ joined #openstack-rally
15:05 yfried amaretskiy: "rally task start" is using the files in "site-packages" and not from the local git repo, so I'm constantly having to run ./install_rally and recreate my deployment to test my work
15:06 yfried amaretskiy: seems wrong to me. Am I doing something wrong?
15:07 amaretskiy yfried there are several ways to develop python package (rally is just python package) and i personally prefer to export PYTHONPATH="path/to/rally"
15:07 yfried amaretskiy: which means?
15:08 amaretskiy https://docs.python.org/2/using/cmdline.html#envvar-PYTHONPATH
15:09 amaretskiy PYTHONPATH env variable allows you load another python code (for example, your /home/username/develop/rally/ instead of /user/lib/python/site-packges/rally)
15:10 amaretskiy that is not a rally question, that is a question "how to run customized python package from another location"
15:12 yfried amaretskiy: ok. btw - network context works great
15:12 yfried amaretskiy: no I have another problem
15:13 yfried amaretskiy: https://review.openstack.org/#/c/143849/1/rally/benchmark/scenarios/nova/utils.py,cm
15:14 yfried amaretskiy: I need to pick the fixed IP of the server, but I don't know the name of the network created by the network context
15:14 amaretskiy server.networks
15:15 yfried amaretskiy: ^ gives a dict where keys are network names
15:16 amaretskiy you should find the way to obtain net name
15:16 amaretskiy but try to avoid api requests
15:17 amaretskiy sorry, server.networks keys are already names
15:18 amaretskiy example: {u'rally_ctx_net_qeybopqncwscbvlv': [u'100.1.0.2']}
15:19 yfried amaretskiy: I know, but "rally_ctx_net_qeybopqncwscbvlv" is a random name, I can't pull it
15:19 amaretskiy the problem is not clear for me
15:19 yfried amaretskiy: I need to get the fixed_ip of the VM
15:20 yfried list_of_fixed_ips = server.networks['<unknown-net-name>']
15:21 yfried so how to I get <unknown-net-name>? the VM could be connected to multiple networks
15:23 amaretskiy fixed_ip = [ip for ip in server.addresses[fixed_network] if ip["version"] == 4][0]["addr"]
15:24 exploreshaifali joined #openstack-rally
15:24 amaretskiy maybe iterate for fixed_network in server.networks.keys(0
15:24 amaretskiy i hope that will work
15:24 amaretskiy let me make the review
15:25 amaretskiy i hope i review your patch today or at tomorrow morning
15:26 yfried amaretskiy: no rush
15:26 amaretskiy :)
15:26 yfried amaretskiy: are you leaving for xmas vacation?
15:26 amaretskiy no
15:27 openstackgerrit Boris Pavlovic proposed stackforge/rally: Refactor validation (part 1)  https://review.openstack.org/143743
15:27 yfried amaretskiy: me neither, so there really is no rush
15:27 amaretskiy okay :)
15:33 yfried amaretskiy: can the network context add router and default GW as well?
16:01 yfried boris-42: amaretskiy: don't know if anyone is interested, but I managed to execute rally from IDE (Pycharm) and use debugger
16:02 amaretskiy yfried: each developer is free to use own trics, for example I use emacs :)
16:02 amaretskiy s/trics/tricks/
16:03 yfried yeah, but redixin said he didn't think it can be done
16:18 openstackgerrit Oleh Anufriiev proposed stackforge/rally: Refactored task report method  https://review.openstack.org/143866
16:29 openstackgerrit Yair Fried proposed stackforge/rally: Fix router GW creation  https://review.openstack.org/143868
16:29 yfried amaretskiy: ^ this is actually a bug in the router creation. simple fix though
16:32 openstackgerrit Oleh Anufriiev proposed stackforge/rally: Refactored task report method  https://review.openstack.org/143866
16:58 jjmb joined #openstack-rally
17:10 rdas joined #openstack-rally
17:15 jjmb joined #openstack-rally
17:31 rvasilets joined #openstack-rally
19:11 yfried joined #openstack-rally
19:19 openstackgerrit joined #openstack-rally
20:07 yfried_ joined #openstack-rally
20:12 yfried__ joined #openstack-rally
22:44 andreykurilin_ joined #openstack-rally
23:34 boris-42 joined #openstack-rally
23:35 tkammer joined #openstack-rally

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