Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2014-11-10

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

All times shown according to UTC.

Time Nick Message
00:08 shakamunyi_ joined #openstack-rally
00:14 shakamunyi_ joined #openstack-rally
00:18 Viswanath joined #openstack-rally
00:34 yaguang joined #openstack-rally
00:44 jjmb joined #openstack-rally
01:19 AlexF joined #openstack-rally
01:36 viswanathsomanch joined #openstack-rally
01:37 viswanathsomanch joined #openstack-rally
01:38 viswanathsomanch joined #openstack-rally
02:13 shakamunyi joined #openstack-rally
02:47 ilbot3 joined #openstack-rally
02:47 Topic for #openstack-rally is now ☁ Rally IRC chat logs http://irclog.perlgeek.de/openstack-rally ☁ Key persons to ask:  boris-42, rediskin, andreykurilin, k4n0  ☁ Open tasks https://trello.com/b/DoD8aeZy/rally ☁ To publish changes to Rally:  https://wiki.openstack.org/wiki/Rally/Develop#How_to_contribute   ☁ Rally weekly updates:  https://wiki.openstack.org/wiki/Rally/Updates
02:49 erkules_ joined #openstack-rally
02:53 Poornima joined #openstack-rally
03:37 jjmb joined #openstack-rally
04:07 viswanathsomanch joined #openstack-rally
04:27 nkhare joined #openstack-rally
04:44 k4n0 joined #openstack-rally
04:50 rdas joined #openstack-rally
04:57 Poornima joined #openstack-rally
05:41 AlexF joined #openstack-rally
05:46 neeti joined #openstack-rally
05:51 chandankumar joined #openstack-rally
06:16 chandan_kumar joined #openstack-rally
06:20 AlexF joined #openstack-rally
06:34 fandi joined #openstack-rally
07:13 AlexF joined #openstack-rally
07:33 AlexF joined #openstack-rally
07:40 jjmb joined #openstack-rally
07:46 tnurlygayanov joined #openstack-rally
08:27 boris-42 joined #openstack-rally
08:29 erwan_taf joined #openstack-rally
08:36 boris-42 erwan_taf: hi there
08:40 AlexF joined #openstack-rally
08:41 erwan_taf hey boris-42
08:52 frobware joined #openstack-rally
09:05 chandan_kumar joined #openstack-rally
09:11 boris-42 erwan_taf: so what you would like to share ?
09:12 boris-42 erwan_taf: do you have some code to share?
09:15 amaretskiy joined #openstack-rally
09:26 oanufriev joined #openstack-rally
09:39 erwan_taf boris-42: yes I do
09:39 erwan_taf boris-42: all the work I've done was developped out of the Ironic context
09:40 erwan_taf boris-42: so take this work as an external tooling that surely needs some adjustements to match Ironic
09:40 erwan_taf boris-42: documentation that explain ideas & concepts, the way it works and some samples is here : https://github.com/enovance/edeploy/blob/master/docs/AHC.rst
09:41 erwan_taf boris-42: the code is everything around health_ in this directory (https://github.com/enovance/edeploy/tree/master/src)
09:41 erwan_taf boris-42: {D}AHC is part of eDeploy up to now
09:42 boris-42 erwan_taf: thanks for links
09:42 boris-42 erwan_taf:  I'll try to analyze them
09:46 ehaselwanter joined #openstack-rally
10:02 vaidy__ril joined #openstack-rally
10:32 erwan_taf joined #openstack-rally
10:40 dmellado joined #openstack-rally
10:54 AlexF joined #openstack-rally
10:58 dmellado joined #openstack-rally
11:21 msdubov joined #openstack-rally
11:24 AlexF joined #openstack-rally
11:28 cdent joined #openstack-rally
11:38 dmellado joined #openstack-rally
11:44 chandan_kumar joined #openstack-rally
12:19 ehaselwanter joined #openstack-rally
12:30 erwan_taf joined #openstack-rally
12:33 ehaselwanter hi there. I got a question regarding predefined usrers https://blueprints.launchpad.net/rally/+spec/benchmark-context-predefined-users
12:33 ehaselwanter would like to run rally without admin rights.
12:35 ehaselwanter or I rephrase my question: Is there kind of an guarantee that running rally with admin rights will not harm a *production* deployment?
12:36 ehaselwanter and the other question is regarding zones. is it possible to specify a zone in the configs?
12:39 pboros joined #openstack-rally
12:41 pboros joined #openstack-rally
12:50 rook "Is there kind of an guarantee that running rally with admin rights will not harm a *production* deployment?" -why would you run Rally against a prodution env?
13:23 amaretskiy rook, there seems to be no guarantee that openstack will break under high load, and another point of risk is resources cleanup - as for cleanup, we have patch which improves safety - https://review.openstack.org/#/c/129060/
13:24 amaretskiy so, I think, there is always a bit of risk if we run high load benchmarks on production
13:24 rook amaretskiy still a very risky recommendation.
13:25 rook For example, of their messaging bus if overloaded, it could break/loose messages due to Rally + their production use.
13:25 rook s/if/is
13:26 rook IMHO I would highly suggest NOT running rally against a production cloud. just my opinion.
13:26 amaretskiy rook, yes. there are several real cases when high load (caused by rally scenarios) even break openstack
13:26 rook I have broken OpenStack many times due to Rally =) isn't that the point?
13:27 amaretskiy rook, personally I think that running on production is not completely safe
13:27 rook Find the maximums of a Cloud ;)
13:27 amaretskiy :)
13:28 rook amaretskiy speaking of breaking... https://gist.github.com/anonymous/4940b19a58a771c68530
13:28 rook https://bugs.launchpad.net/rally/+bug/1390993
13:29 rook ever seen that running Rally?
13:30 amaretskiy rook, no, I saw another issues related to broken keystone
13:30 rook interesting, I haven't seen keystone issues
13:31 amaretskiy rook, that is a point of benchmark type and arguments, and keystone version :)
13:31 rook also, i haven't ran keystone scenario's
13:31 rook i have focused on vmtasks work.
13:43 ehaselwanter @rook well. I would love to leverage the encoded "verification" part of rally ...
13:43 rook isn't that a tempest run?
13:43 * rook hasn't ran verification...
13:44 rook also, aren't you putting the cart before the horse? running verification after you make it a production env?
13:44 ehaselwanter rally has definded some use cases. start/stop .. well, use cases
13:44 ehaselwanter if I run it is see that they work
13:45 ehaselwanter I can re-implement that, but why?
13:46 ehaselwanter and I'm thinking POC not production ... that is the *production* in stars
13:46 ehaselwanter where you have already users
13:46 ehaselwanter but can do some load testing ...
13:47 ehaselwanter so. I just wanted to know if I can run rally and the cluster is not broken afterwards. or "disable this test and you are fine" ... kind of advice
13:47 ehaselwanter what I hear now is: don't run it ... is that correct?
13:48 boris-42 joined #openstack-rally
13:49 amaretskiy ehaselwanter, `rally verify' is actually tempest run
13:50 amaretskiy ehaselwanter, there is no way to be sure that production cluster is fine after rally scenario
13:51 amaretskiy ehaselwanter, because 1) patch that improves cleanup is not merged yet, and 2) high load can break the cluster
13:51 ehaselwanter @amaretskiy why is that?
13:51 amaretskiy ehaselwanter, please make the question more clear
13:51 ehaselwanter @amaretskiy if the predefined user blueprint is implemented and I do rate limiting, it should be fine, right?
13:52 amaretskiy ehaselwanter, if your benchmarks make low load then you will probably not break the cluster, and that is ok :)
13:53 amaretskiy ehaselwanter, that is usually a point of cluster perfomance
13:54 boris-42 ehaselwanter:  hi there
13:55 amaretskiy ehaselwanter, you have asked about re-implementation of use cases - you can write your own benchmark scenarios, or use existent scenarios (doc/samples/...) with different arguments
13:55 ehaselwanter hm. so I should prepare a stress-testing window? and "repair" the cluster after running rally?
13:55 boris-42 ehaselwanter: so we are working hard to make cleanup process really "production" ready
13:55 boris-42 ehaselwanter: like if something went wrong with rally in middle of benchmarking and process dies
13:55 ehaselwanter boris-42: this is confusing now
13:55 boris-42 ehaselwanter: what is confusing?)
13:56 ehaselwanter just heard I never should run rally against production
13:56 ehaselwanter and you telling me I *will* be able to do that (what I expected)
13:56 boris-42 ehaselwanter:  hm not sure that I get, for now in cleanup process there is no retries
13:56 boris-42 ehaselwanter:  they will be
13:57 boris-42 ehaselwanter:  I think they will be tomorrow actually
13:57 ehaselwanter tomorrow?
13:57 boris-42 ehaselwanter:  ya patches are ready
13:57 ehaselwanter for the predefined user?
13:57 ehaselwanter or cleanup?
13:57 boris-42 ehaselwanter:  cleanup*
13:58 ehaselwanter what about [Benchmarking with predefined users](https://blueprints.launchpad.net/rally/+spec/benchmark-context-predefined-users)
13:58 ehaselwanter what is blocking this blueprint?
13:58 boris-42 ehaselwanter:  so I am working on it as well...
13:58 ehaselwanter would really love to create a benchmarking sandbox (aka create tenant and user)
13:58 boris-42 ehaselwanter: there are few small blockers, I am refactoring actually validation process
13:58 boris-42 ehaselwanter:  yep deffently understand you
13:59 ehaselwanter and not provide the admin creds
13:59 boris-42 ehaselwanter:  yep yep
13:59 boris-42 ehaselwanter:  so we were refactroing a lot rally to do that
13:59 ehaselwanter just to prevent unwanted side effekts
13:59 ehaselwanter just to prevent unwanted side effects
13:59 boris-42 ehaselwanter:  there is very last step to do
13:59 ehaselwanter like delete all networks
13:59 boris-42 ehaselwanter:  just create one more context
13:59 boris-42 ehaselwanter:  + a bit imporve validation process
13:59 boris-42 ehaselwanter:  this should be done as well soon
14:00 ehaselwanter cool.
14:00 boris-42 ehaselwanter:  in any case I will publish to mailing list some announces
14:00 ehaselwanter I know. I hate to ask the question (and not contributing a thing) ... is there an ETA for the user thing?
14:00 ehaselwanter like *maybe this month* ?
14:00 boris-42 ehaselwanter: I hope it will be faster
14:00 ehaselwanter :-)
14:01 boris-42 ehaselwanter:  cause there is very last block
14:01 ehaselwanter cool
14:01 boris-42 ehaselwanter:  just new context + a bit imporve validation
14:01 ehaselwanter antoher thing: can I specify a default zone?
14:01 boris-42 ehaselwanter:  so we are supporting regions
14:01 boris-42 ehaselwanter:  but as far as I know default zones are admin only stuff?
14:02 boris-42 ehaselwanter:  or no?
14:02 ehaselwanter to boot an instance?
14:04 ehaselwanter nova boot -availability-zone myzone
14:04 ehaselwanter ?
14:04 openstackgerrit joined #openstack-rally
14:04 ehaselwanter no. that is not a admin thing
14:05 ehaselwanter boris-42: it is an admin thing to specify which hosts are in ther
14:05 ehaselwanter boris-42: it is an admin thing to specify which hosts are in there
14:05 ehaselwanter but I would like to specify the boot parameter
14:06 ehaselwanter typically it is RegionOne .. or something like that
14:14 mwagner_lap joined #openstack-rally
14:15 boris-42 ehaselwanter:  so
14:16 boris-42 ehaselwanter:  https://github.com/stackforge/rally/blob/master/doc/samples/deployments/existing.json#L4
14:16 boris-42 ehaselwanter: does that works for you ^
14:17 jjmb joined #openstack-rally
14:18 chandankumar joined #openstack-rally
14:23 boris-42 erwan_taf:  so it's something like ironic?
14:23 boris-42 erwan_taf:  I mean edeploy
14:24 dmellado joined #openstack-rally
14:30 ehaselwanter boris-42: thx
14:35 chandankumar joined #openstack-rally
14:37 jjmb joined #openstack-rally
14:38 erwan_taf boris-42: on some part yes
14:39 erwan_taf boris-42: it does got a little bit further especially for initial deployment
14:39 boris-42 erwan_taf: so what can i do with that ? deploy OpenStack ?
14:39 boris-42 erwan_taf: or just make provision?
14:40 erwan_taf install any OS based on an hardware description
14:40 boris-42 erwan_taf:  so we can make nice "serverprovider" on top of it in rally?)
14:46 erwan_taf I'm not sure to understand this "superprovider" word
14:49 boris-42 erwan_taf: serverprovider
14:50 boris-42 erwan_taf: "rally deploy" commands deploys openstack clouds
14:50 boris-42 erwan_taf:  combining other tools
14:50 erwan_taf edeploy alone doesn't provide a working cloud
14:51 erwan_taf it only do the installation + basic configuration (network, bootloader, ..)
14:51 erwan_taf then we use puppet & so to do the configuration to gain a cloud
14:51 boris-42 erwan_taf: in rally we have 2 parts
14:52 boris-42 erwan_taf:  provision + deploying
14:52 boris-42 erwan_taf:  so provision part can be based on edeploy*
14:52 boris-42 erwan_taf:  that is what I was talking about =)
14:52 erwan_taf ok sorry, I don't know rally very well yet :-
14:52 erwan_taf :(
14:54 boris-42 erwan_taf: no worries
14:54 boris-42 erwan_taf:  so I didn't get fully what did you want to share with us?
14:55 boris-42 erwan_taf:  there are some benchmarks in edeploy?
14:55 boris-42 erwan_taf:  or?
15:00 erwan_taf yes there is
15:00 erwan_taf in srC/
15:01 erwan_taf eDeploy have a feature to build reproducible operating system
15:01 erwan_taf eDeploy can build my benchmark tool inside one which is called health
15:24 openstackgerrit Sergey Skripnick proposed a change to stackforge/rally: Add Dockerfile  https://review.openstack.org/132556
15:30 chandankumar joined #openstack-rally
15:35 klindgren joined #openstack-rally
15:37 msdubov joined #openstack-rally
15:40 jjmb1 joined #openstack-rally
15:40 erwan_taf joined #openstack-rally
15:47 Viswanath joined #openstack-rally
15:51 chandankumar joined #openstack-rally
15:51 jjmb joined #openstack-rally
15:57 Viswanath joined #openstack-rally
16:08 Viswanath joined #openstack-rally
16:22 dmellado joined #openstack-rally
16:35 dmellado joined #openstack-rally
16:39 dmellado joined #openstack-rally
16:44 marcoemorais joined #openstack-rally
16:45 msdubov joined #openstack-rally
16:53 jjmb1 joined #openstack-rally
17:05 dmellado joined #openstack-rally
17:13 AlexF joined #openstack-rally
17:19 dmellado joined #openstack-rally
17:21 dmellado_ joined #openstack-rally
17:29 Viswanath joined #openstack-rally
17:29 chandankumar joined #openstack-rally
17:30 marcoemorais joined #openstack-rally
17:31 marcoemorais joined #openstack-rally
17:36 marcoemorais joined #openstack-rally
17:38 marcoemorais joined #openstack-rally
17:41 chandan_kumar joined #openstack-rally
17:44 openstackgerrit Alexander Maretskiy proposed a change to stackforge/rally: Base HTML templates and improvements for task report  https://review.openstack.org/131844
17:47 Viswanath joined #openstack-rally
17:49 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 3  https://review.openstack.org/129060
17:49 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 2  https://review.openstack.org/132437
17:49 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 1  https://review.openstack.org/132436
18:02 Viswanath joined #openstack-rally
18:05 marcoemorais joined #openstack-rally
18:09 marcoemorais joined #openstack-rally
18:09 marcoemorais joined #openstack-rally
18:15 Viswanath joined #openstack-rally
18:21 marcoemorais joined #openstack-rally
18:24 marcoemorais joined #openstack-rally
18:28 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 3  https://review.openstack.org/129060
18:28 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 2  https://review.openstack.org/132437
18:28 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 1  https://review.openstack.org/132436
18:35 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 3  https://review.openstack.org/129060
18:35 openstackgerrit Boris Pavlovic proposed a change to stackforge/rally: Refactor generic cleanup mechanism part 2  https://review.openstack.org/132437
18:42 Viswanath joined #openstack-rally
18:47 fandi joined #openstack-rally
18:48 Viswanath joined #openstack-rally
18:57 vhoskaot joined #openstack-rally
18:59 vhoskaot Hi friends, is there any way in rally to create 100 networks each with 10 VMs.. how can I mention this in the json file --> Create 100 networks, attach 10 VMs to each network
19:01 vhoskaot can anyone please help me...
19:03 vhoskaot hello!
19:05 vhoskaot anybody there!
19:09 vhoskaot h e l l o !!
19:11 rediskin joined #openstack-rally
19:12 vhoskaot can anyone please help me ..
19:12 Viswanath joined #openstack-rally
19:14 jjmb joined #openstack-rally
19:14 vhoskaot is there any way in rally to create 100 networks each with 10 VMs.. how can I mention this in the json file --> Create 100 networks, attach 10 VMs to each network
19:18 boris-42 vhoskaot:  i have flight in 3 minutes
19:18 boris-42 rediskin: please help vhoskaot
19:19 rediskin boris-42: i have some problems with network
19:19 boris-42 rediskin: at least irc work
19:20 boris-42 rediskin: so help please vhoskaot
19:21 rediskin not sure if it possible
19:21 vhoskaot is it not possible to attach VMs to networks for Rally json ?
19:22 vhoskaot boris-42: happy journey
19:22 boris-42 rediskin: why not?
19:22 rediskin boris-42: is networks context merged?
19:23 boris-42 rediskin:  nope but he can fetch it
19:23 vhoskaot help me friends
19:23 rediskin it is possible to create 100 networks, and it is possible to create vm's
19:23 rediskin but in separate taks
19:23 boris-42 rediskin: i don't have laptoop at moment so rediskin please
19:24 rediskin vhoskaot: you could try to fetch that patch
19:25 vhoskaot rediskin: which patch ? I have git cloned rally
19:27 jjmb joined #openstack-rally
19:27 Viswanath joined #openstack-rally
19:27 rediskin vhoskaot: oh sorry, that patch isn't working yet
19:27 rediskin https://review.openstack.org/#/c/103306/70/rally/benchmark/context/networks.py
19:28 rediskin here it whole patch https://review.openstack.org/#/c/103306/70
19:28 rediskin is*
19:28 rediskin vhoskaot: this patch will be fixed and merged soon
19:29 vhoskaot rediskin: cool, thanks a lot
19:29 vhoskaot rediskin: does this patch allow us to attack VMs to networks in the json file ?
19:29 vhoskaot attach*
19:30 rediskin It allows us to create N networks before running any scenario
19:30 rediskin Not sure about attaching
19:31 rediskin AFAIR nova scenarios allows us to attach booted VM's to random networks
19:31 rediskin So it will be possible to do almost the same that you need
19:31 vhoskaot rediskin: Here is what I want to do... create 100 networks with 10 VMs attached to each network..
19:32 rediskin I got it. Seems one more patch needed to do this
19:33 vhoskaot rediskin: what networks do nova scenarios use.. for e.g., NovaServers.boot_and_delete_server
19:34 rediskin vhoskaot: it is valuable scenario. we definitely should make this possible
19:34 vhoskaot rediskin: can I file a bug for it ?
19:34 rediskin vhoskaot: it would be nice
19:35 rediskin currenly nova scenario allows to specify one network for all booted VMs
19:36 vhoskaot rediskin: ok, who creates that "one" network
19:37 rediskin that one network assumed is already created
19:37 vhoskaot rediskin: https://bugs.launchpad.net/rally.. is this the right place to file the bug ?
19:37 rediskin yes
19:38 vhoskaot rediskin: thanks
19:38 vhoskaot rediskin: already created.. meaning the ones that show up in "rally show networks"
19:39 vhoskaot rediskin: already created.. meaning the ones that show up in "rally show networks" ?
19:39 rediskin vhoskaot: i dont get it. launchad is suggesting that bug already reported?
19:40 rediskin oh, got it
19:40 vhoskaot rediskin: which bug ? I have not filed any bug yet
19:40 rediskin yes, one in "rally show networks"
19:40 vhoskaot rediskin: can you send the bug info if there exists one..
19:40 rediskin are you using nova network?
19:40 vhoskaot rediskin: no, I am using neutron
19:41 vhoskaot rediskin: can you send the bug info if there exists one..
19:41 rediskin vhoskaot: no, i though "already created" is about bug
19:41 rediskin but you talking about network
19:42 rediskin so, yes. nova scenario will attach all vms to one specified network
19:42 vhoskaot rediskin: no no... My use case is to create X number of networks and attach Y number of VMs per network, and boot all the X*Y VMs... Since we cannot do this in rally today, I will a bug for this
19:43 vhoskaot rediskin: the json need to take a range to network IDs and have a parameter like may be VMs_per_neywork...
19:44 vhoskaot rediskin: Am I making sense ?
19:44 rediskin it may be one scenario kind of "Neutron.create_networks_and_attach_vms" with arguments "number_of_networks" and "vms_per_network"
19:45 vhoskaot rediskin: that is the exact scenario I am talking about.. don't you think it is a nice scenario that will be useful ?
19:45 rediskin I think its hard to do other solution
19:46 vhoskaot rediskin: which solution ? "Neutron.create_networks_and_attach_vms" with arguments "number_of_networks" and "vms_per_network" ?
19:46 rediskin no, other then this
19:46 rediskin I think it is most simple way
19:46 vhoskaot rediskin: righ right, "Neutron.create_networks_and_attach_vms" with arguments "number_of_networks" and "vms_per_network" is good enough.. no other solution is needed
19:47 vhoskaot rediskin: but, the user must make sure to delete the VMs and networks himself/herself elsewhere like for NovaServers.boot_server case
19:47 marcoemorais joined #openstack-rally
19:47 vhoskaot rediskin: shall I go ahead and file the bug then ?
19:49 rediskin vhoskaot: sure
19:49 vhoskaot rediskin: thanks a lot for your help !
19:51 Viswanath joined #openstack-rally
19:51 rediskin vhoskaot: you welcome. also, maybe somebody propose better solution in comments =)
19:52 vhoskaot rediskin: another question. Is there anyway to enable dhcp for nova boot case ?
19:53 mcaulfie joined #openstack-rally
19:54 rediskin vhoskaot: how? by modifying existing network?
19:54 rediskin we should edit network to enable dhcp, right?
19:56 vhoskaot rediskin: there is nothing is the json file that connects a port (port ID) and a network (network ID) to the vm that is booted.. so, I'm wondering how we can give dhcp option for nova boor scenario
19:57 vhoskaot like for e.g., can I give something like enable_dhcp = true under "args" in NovaServers.boot_server ?
19:59 rediskin vhoskaot: argument given for scenarios NovaServers.boot_ will passed to nova-pythonclient
19:59 rediskin you may specify any argument for e.g. NovaServers.boot_server
20:01 rediskin and this arguments will be passed to novaclient. if novaclient accepts enable_dhcp so you could specify it
20:03 vhoskaot rediskin: No nova does not understand dhcp... the earlier nova-network unserdtand dhcp option.. I am not using nova-network.. I am using neutron.. so, I can specify dhcp either in neutron's create-subnet or attach a neutron network by using "neutron dhcp-agent-network-add "
20:03 vhoskaot rediskin: and both are invisible to nova boot scenarios for rally's json...
20:04 rediskin so, we need to modify existing network, or precreate network with dhcp enabled
20:04 vhoskaot rediskin: yes
20:05 rediskin second option will be done by networkContext (that patch which will be fixed and merged soon)
20:06 vhoskaot rediskin: going back to the case "Neutron.create_networks_and_attach_vms" with arguments "number_of_networks" and "vms_per_network"... there must also be another option like "enable_dhcp_for_all_networks"...
20:07 rediskin Why? We can create networks with any options
20:07 rediskin in our first case
20:07 rediskin (Neutron.create_networks_and_attach_vms)
20:07 vhoskaot rediskin: right, thanks for the patch that precreates network with dhcp
20:08 rediskin we can pass any argument to create_networks_ part, and create network with or without dhcp
20:08 rediskin and boot and attach vms then
20:08 vhoskaot rediskin: yes, the create_network part can handle the dhcp part.. agreed
20:08 vhoskaot rediskin: cool, thanks for your help.. will file the bug soon
20:09 rediskin you welcome
20:10 marcoemorais joined #openstack-rally
20:23 dmellado_ joined #openstack-rally
20:38 afazekas joined #openstack-rally
21:22 dmellado_ joined #openstack-rally
21:30 vhosakot joined #openstack-rally
21:32 vhosakot hi, I added a new scenario and a json... and I see these errors when I run the new scenario.. can anyone tell me what I'm missing...
21:32 vhosakot <class 'rally.exceptions.NotFoundScenarios'>
21:32 vhosakot There are no benchmark scenarios with names: `NeutronNetworks.create_networks_and_boot_server`.
21:35 boris-42 vhosakot: hey
21:35 vhosakot boris-42: hi
21:35 boris-42 vhosakot:  so basically maybe you have some mistakes in module
21:35 boris-42 vhosakot:  with your benchmark
21:36 boris-42 vhosakot:  to test that just do something like rally -i -m full.name.to.the.module.with.benchmark
21:36 boris-42 python -i -m**
21:36 vhosakot this is throwing the error  -  "rally -v task start rally/doc/samples/tasks/scenarios/neutron/create_network_and_boot_servers.json"
21:37 boris-42 vhosakot:  it is not thorowing error..
21:38 vhosakot <class 'rally.exceptions.NotFoundScenarios'>
21:38 boris-42 vhosakot:  it just validation of rally task
21:38 vhosakot There are no benchmark scenarios with names: `NeutronNetworks.create_networks_and_boot_server`.
21:38 boris-42 vhosakot:  I think this is pretty good output for validation failure
21:38 boris-42 vhosakot:  so basically it says no such class + method
21:39 boris-42 vhosakot:  so it means that you have mistakes somewhere in description
21:39 vhosakot I do have the method def create_networks_and_boot_server in rally/benchmark/scenarios/neutron/network.py
21:39 boris-42 vhosakot:  as I said before it's simple to test... just try to import module with benchmark
21:40 boris-42 vhosakot:  so please make "python -i -m rally.benchmark.scenarios.neutron.network"
21:40 vhosakot boris-42: you mean "python -i -m rally/benchmark/scenarios/neutron/network.py" ?
21:40 boris-42 vhosakot:  ya
21:40 vhosakot localadmin@ubuntu-14-4:~/rally$ python -i -m rally/benchmark/scenarios/neutron/network.py
21:40 vhosakot Traceback (most recent call last):
21:40 vhosakot File "/usr/lib/python2.7/runpy.py", line 156, in _run_module_as_main
21:40 vhosakot sys.exit(msg)
21:40 vhosakot SystemExit: /usr/bin/python: Import by filename is not supported.
21:40 vhosakot >>>
21:40 vhosakot here are all the lines in the output
21:41 boris-42 no no
21:41 boris-42 don't write .py
21:41 boris-42 just
21:41 boris-42 python -i -m rally.benchmark.scenarios.neutron.network
21:41 boris-42 -m is similar to (import rally.benchmark.scenarios.neutron.network in code)
21:41 boris-42 just one step shorter
21:42 vhosakot boris-42: I see no error, and go to the python prompt >>>
21:42 boris-42 vhosakot:  nice
21:42 boris-42 vhosakot:  then paste please your code
21:42 vhosakot sure.. 1 sec
21:42 boris-42 vhosakot:  and be quite fast=)
21:42 boris-42 vhosakot:  cause next flight will be soon
21:46 boris-42 vhosakot:  seems like I lost connection
21:46 boris-42 vhosakot:  could you resend please*
21:47 vhosakot boris-42: here is all the code  -  http://fpaste.org/149494/14156560/
21:49 boris-42 hm
21:49 boris-42 vhosakot: did you reinstall rally?
21:49 boris-42 vhosakot:  I mean running "sudo python setup.py install"
21:50 boris-42 vhosakot:  if you don't want to reinstall it every time
21:50 boris-42 you can put plugins
21:50 boris-42 inside "~/.rally/plugins"
21:50 boris-42 vhosakot:  really sorry I have to go =)
21:51 Viswanath_ joined #openstack-rally
21:51 Viswanath joined #openstack-rally
21:54 boris-42 vhosakot: have a bit more time
21:54 boris-42 vhosakot:  so did resintall help?
21:55 vhosakot boris-42: should I reinstall rally when I add a new scenario / method in NeutronNetworks class ?
21:56 boris-42 vhosakot: if you change code you need
21:56 boris-42 vhosakot:  except if you are using plugins
21:56 boris-42 vhosakot:  and putting everything to "~/.rally/plugins/"
21:56 boris-42 vhosakot:  then you don't need to reinstall
21:57 boris-42 vhosakot:  1% charge =(
22:01 vhosakot boris-42: still same error even after reinstalling rally
22:10 rediskin left #openstack-rally
22:25 Viswanath joined #openstack-rally
22:36 dmellado_ joined #openstack-rally
22:41 leeantho joined #openstack-rally
23:08 vhosakot joined #openstack-rally
23:39 openstackgerrit Anthony Lee proposed a change to stackforge/rally: Add a servers context and a cinder stress scenario  https://review.openstack.org/127392

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