Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel-dev, 2015-08-20

| Channels | #fuel-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 mkwiek joined #fuel-dev
00:27 mkwiek joined #fuel-dev
00:30 dims joined #fuel-dev
01:41 xarses joined #fuel-dev
02:21 dims joined #fuel-dev
02:30 dims joined #fuel-dev
03:04 gongysh joined #fuel-dev
05:31 neophy joined #fuel-dev
05:58 mkwiek joined #fuel-dev
06:18 neophy joined #fuel-dev
06:25 mkwiek joined #fuel-dev
06:39 stamak joined #fuel-dev
06:47 Hulda joined #fuel-dev
06:49 mkwiek joined #fuel-dev
06:58 dkusidlo joined #fuel-dev
07:10 amnk joined #fuel-dev
07:18 yottatsa joined #fuel-dev
07:21 amnk joined #fuel-dev
07:34 yottatsa joined #fuel-dev
07:56 hyperbaba joined #fuel-dev
07:58 stamak joined #fuel-dev
08:02 romcheg joined #fuel-dev
08:08 romcheg jaypipes: Hi, could you please poke back when you're available?
08:25 romcheg left #fuel-dev
08:26 romcheg joined #fuel-dev
08:27 romcheg bookwar: hi, around?
08:28 ekosareva joined #fuel-dev
08:55 ashtokol_ joined #fuel-dev
08:56 ashtokol_ joined #fuel-dev
09:09 bookwar romcheg: yes
09:10 romcheg bookwar: I've got a question about aborting jobs
09:11 romcheg Yesterday I caused one job to be aborted and that left nailgun running
09:11 romcheg so my question is how do jobs get aborted?
09:13 romcheg I mean, there is a shell script which is creating a virtualenv and starting tox
09:13 romcheg or run_test.sh
09:13 romcheg I can see nothing there that can process any interuptions
09:19 romcheg bookwar: so is that a SIGINT or any other signal which is sent to the shell script?
09:25 neophy joined #fuel-dev
09:26 stamak joined #fuel-dev
09:29 stamak joined #fuel-dev
09:31 stamak joined #fuel-dev
09:33 SergK romcheg, are you talking about this? https://github.com/stackforge/fuel-web/blob/master/run_tests.sh#L561-L575
09:34 romcheg SergK: no, about the builder script
09:34 romcheg Which for me is fuel-ci/builders/verify-python-fuelclient.sh
09:38 romcheg SergK: in Fuel Client we don't have run_test.sh anymore. We killed it with fire :)
09:38 yottatsa joined #fuel-dev
09:38 monester joined #fuel-dev
09:39 degorenko joined #fuel-dev
09:49 yottatsa joined #fuel-dev
09:55 yottatsa joined #fuel-dev
09:56 dims joined #fuel-dev
10:36 ashtokol_ joined #fuel-dev
10:57 ddmitriev joined #fuel-dev
11:10 dstepanenko joined #fuel-dev
11:16 dims joined #fuel-dev
11:21 yottatsa joined #fuel-dev
11:22 yottatsa_ joined #fuel-dev
11:33 dims joined #fuel-dev
11:34 dkusidlo joined #fuel-dev
11:35 yottatsa joined #fuel-dev
11:46 amnk joined #fuel-dev
11:52 jaypipes romcheg: hi!
11:53 romcheg jaypipes: Hi, I have a question regarding OpenStack CI
11:53 jaypipes romcheg: go for it :)
11:53 romcheg Yestedray I finally enabled python jobs for fuel client
11:54 romcheg Now it takes a while for them to pass. Is there a way to speed up openstack gates?
11:57 jaypipes romcheg: what is "a while"? :) can you give me a link to an example that takes a while?
11:59 hatuomin joined #fuel-dev
11:59 romcheg A while is 30-40 minutes
11:59 romcheg Usually it took 10
12:00 romcheg I know it's ok in the upstream world but I foresee Mirantis folks complaining :)
12:01 jaypipes romcheg: Python jobs for upstream Nova (15K+ unit and functional tests) take only around 6 minutes :)
12:02 jaypipes romcheg: do you have a link to an examlpe? is it the actual tests that take that long, or the build ofthe slave instance for testing?
12:03 romcheg jaypipes: The tests themselves take like a minute. Waiting in a queue takes a while
12:03 romcheg jaypipes: https://review.openstack.org/#/c/215077/
12:03 romcheg You can check ^
12:03 * jaypipes looks :)
12:06 jaypipes romcheg: ya, unit tests took about 2.5 minutes... is there a way to see the logs on fuel-infra.org's side for this? "Building remotely on ci-slave15.fuel-infra.org"
12:07 romcheg jaypipes: on fuel infra they unit an integrational tests run at the same time
12:07 romcheg jaypipes: It's fine that tests take 2.5 minutes, I'm worrying more about a long queue when there are not enough workers on openstack ci
12:07 jaypipes romcheg: no, sorry, I should rephrase my question.... I meant, I'm trying to determine how long it took for the upstream CI to "grab fuel-infra node #15"
12:09 romcheg jaypipes: you have more expertise in this than I do so I don't know :)
12:11 bookwar romcheg: jaypipes are you talking about Openstack Infra CI or Fuel CI?
12:12 romcheg bookwar: about Openstack Infra CI, The queue there is pretty long, so I'm curious whether there is a way of optimizing smth
12:12 jaypipes romcheg: so, are you *sure* that it's taking a long time? from the review, I see timur committed the patch at 7:55am, that the fuel infra CI system grabbed it at 7:55:42 AM, and that the result was updated back to upstream Gerrit at 7:59am.
12:13 romcheg jaypipes: Yesterday in the evening I had to wait 30-40 minutes
12:13 romcheg lemme find those patches pls
12:13 mkwiek1 joined #fuel-dev
12:13 jaypipes romcheg: k, thanks :)
12:13 bookwar jaypipes: you are looking at the wrong CI user. There are Fuel CI votes made by fuel-infra, and those are ok
12:13 bookwar but there are gate checks made by Jenkins user
12:13 jaypipes I see... if you could point me to an examlpe, that would be great
12:14 bookwar and the 'Jenkins' votes appear when change passes through the Openstack Infra queue
12:14 romcheg jaypipes: https://review.openstack.org/#/c/214674/
12:14 romcheg I was testing the patch
12:14 romcheg *the jobs
12:15 romcheg I posted a patch set on 5:41 and got the score on 6:06
12:15 romcheg so almost half an hour
12:16 jaypipes romcheg: http://status.openstack.org/zuul/
12:17 jaypipes romcheg: looks like you may have just been unfortunate to hit a peak in node usage...
12:17 jaypipes romcheg: look at the graph at bottom of the page
12:18 youellet_ joined #fuel-dev
12:18 jaypipes romcheg: there may have been some large patch series in other projects that pulled a bunch of nodes from nodepool and made the queue longer.
12:18 romcheg jaypipes: yeah, looks like that. But IIRC node usage usually increases on evenings
12:18 jaypipes romcheg: well, no, it goes in cycles based on europe, then us, then asia timezones :)
12:18 romcheg On European evenings :D
12:19 jaypipes romcheg: the launch times are affected by the quotas that the upstream CI users on RAX cloud and HP cloud have, along with the general launch times of those different clouds.
12:19 romcheg jaypipes: btw, what do grey jobs mean on ZUUL status page?
12:19 jaypipes romcheg: hop on to #openstack-infra and ask the folks there what the average time to "grab a slave node for python gate jobs" is.
12:20 jaypipes romcheg: grey jobs are queued with no node.
12:20 jaypipes romcheg: meaning they haven't been assigned a free node yet (AFAIK)
12:20 romcheg thx
12:21 romcheg jaypipes: I'll keep watching on our patches, if it still takes a while for grabbing a node, I'll ask on os-infra
12:21 romcheg thanks for the information :)
12:22 jaypipes romcheg: hey, any time Roman :) it might be nice if you posted a quick email to the ML saying something to the effect of "hi infra team, I'm wondering if you have some data on the average wait times for a Python gate job to get a slave node from nodepool. Is there anything we can help with to reduce the average times?"
12:30 romcheg jaypipes: I'm afraid they will ask to donate some nodes :D
12:30 bookwar and some engineers :)
12:30 jaypipes romcheg: meh, they might, but at least we will have reached out to see if we can help. :)
12:48 dkusidlo joined #fuel-dev
13:00 gongysh joined #fuel-dev
13:02 dkusidlo joined #fuel-dev
13:09 ashtokol_ joined #fuel-dev
13:15 aviramb joined #fuel-dev
13:25 bogdando joined #fuel-dev
13:26 neophy joined #fuel-dev
13:28 dkusidlo joined #fuel-dev
13:35 dims_ joined #fuel-dev
13:46 dims joined #fuel-dev
13:51 dkusidlo joined #fuel-dev
13:58 agordeev joined #fuel-dev
13:58 ylobankov joined #fuel-dev
13:59 degorenko joined #fuel-dev
14:04 warpc__ joined #fuel-dev
14:09 yottatsa joined #fuel-dev
14:11 yottatsa joined #fuel-dev
14:18 yottatsa joined #fuel-dev
14:19 kpimenova joined #fuel-dev
14:32 romcheg joined #fuel-dev
14:43 romcheg1 joined #fuel-dev
14:46 romcheg joined #fuel-dev
14:47 yottatsa joined #fuel-dev
14:59 yottatsa joined #fuel-dev
15:09 yottatsa joined #fuel-dev
15:10 rmoe joined #fuel-dev
15:18 yottatsa joined #fuel-dev
15:51 angdraug joined #fuel-dev
15:53 neophy joined #fuel-dev
15:57 AlexeyElagin joined #fuel-dev
15:58 yottatsa joined #fuel-dev
16:05 skolekonov joined #fuel-dev
16:19 rawat joined #fuel-dev
16:21 rawat Hi everyone, I am developing on plugin which will do some changes on all controller and compute, But in compute I want plugin will execute on only KVM or Qumu hypervise, How can i do this ...??
16:23 rawat can any one please help me ...
16:23 ikalnitsky rawat. hi. i bet i asked you day ago what do you mean by "KVM or Quemu" hypervisor? do you mean *all compute nodes* or what?
16:24 rawat ikalnitsky : yes i given reply but late..
16:24 rawat ikalnitsky : yes all compute node
16:24 ikalnitsky which fpb version do you use?
16:25 rawat package version 2.0.0
16:25 rawat how can i see fpb version
16:26 ikalnitsky if it was installed from PyPI (using pip) then yeah, it's 2.0. currently we have 3.0 only in master branch
16:26 ikalnitsky but ok
16:26 ikalnitsky so, basically you need to add tasks.yaml to your plugin
16:26 ikalnitsky a simple task
16:26 ikalnitsky what problem do you have there? we have a good example, aren't we?
16:26 rawat i had added one condition but it is not working
16:27 ikalnitsky afaiu, it should be something like http://xsnippet.org/360917/
16:27 rawat condition: "settings:common.libvirt_type.value == 'kvm' or settings:common.libvirt_type.value == 'qemu'"
16:28 rawat in my cluster there is one vmware compute node also there..
16:28 rawat plugin should not apply to vmware compute node..
16:30 ikalnitsky this conditions are applied only for UI. it has nothing in common with deployment
16:30 ikalnitsky what is the role of vmware compute node? does it called "compute" too?
16:30 ikalnitsky if so, you need to check this in runtime during execution and skip your deployment code if it's vmware-compute
16:31 rawat http://paste.openstack.org/show/422729/ this is my task.yaml
16:31 rawat yes
16:31 rawat where i need to change..??
16:33 ikalnitsky rawat, hm.. line 12.. why do you think your condition works? afaik, we don't support this one.
16:34 ikalnitsky evgenyl, could you please take a look? ^^
16:34 rawat it's not working..
16:34 ikalnitsky rawat why it should?
16:35 ikalnitsky as i said, you run task on all computes.. and your 'puppet/manifests/compute.pp' should check whether it's quemu/kvm node or not
16:35 ikalnitsky afaik, we don't support other options right now. no conditions.
16:35 evgenyl I thought we don't have UI style conditions for tasks.
16:36 rawat is there any value which can say what is the hypervisor...?
16:38 ikalnitsky value where? in astute.yaml? definetely should be something, otherwise the node won't be deployed with vmware
16:38 rawat yes
16:38 ikalnitsky i can recommend you to check hiera
16:38 ikalnitsky or ask our partner intergration team
16:39 rawat yes if any value is there in astute.yaml file then i can make it..
16:40 ikalnitsky rawat, i think you can even check /etc/nova/nova-compute.conf - there should be an option that said that nove-compute works with vmware now
16:41 rawat git it in astute.yaml file
16:41 rawat libvirt_type: kvm
16:42 rawat hiera libvirt type
16:42 ikalnitsky if hiera('use_vcenter', false) or hiera('libvirt_type') == 'vcenter'
16:42 ikalnitsky that's what i found in our puppet
16:43 rawat yes it wil work for me .. :)
16:45 rawat thank ikalnitsky :)
16:45 monester joined #fuel-dev
16:46 ikalnitsky rawat, you're welcome
16:47 rawat ikalnitsky : can i contact you if i have any question related to plugin ..?
16:49 ikalnitsky yeah, sure. but i'll be in vacation next week, so i will be barely available.
16:50 rawat ikalnitsky : ok thanks a lot.. :) happy vacation
16:50 ikalnitsky thanks =)
17:05 romcheg joined #fuel-dev
17:06 mattymo_ come one, come all, review some code today! https://review.openstack.org/215228
17:08 romcheg1 joined #fuel-dev
17:18 romcheg joined #fuel-dev
17:32 ashtokol_ joined #fuel-dev
17:36 neophy joined #fuel-dev
17:43 yottatsa joined #fuel-dev
17:49 xarses joined #fuel-dev
18:03 jfluhmann joined #fuel-dev
18:39 angdraug joined #fuel-dev
18:41 Samos123 joined #fuel-dev
19:00 Samos123 joined #fuel-dev
19:10 Samos123 joined #fuel-dev
19:19 stamak joined #fuel-dev
19:26 dims_ joined #fuel-dev
19:33 vkramskikh joined #fuel-dev
19:34 rvyalov joined #fuel-dev
19:34 xenolog joined #fuel-dev
19:35 meow-nofer joined #fuel-dev
19:36 sbog joined #fuel-dev
19:41 alex_didenko joined #fuel-dev
19:43 Samos123 joined #fuel-dev
20:04 stamak joined #fuel-dev
20:17 neophy joined #fuel-dev
20:30 e0ne joined #fuel-dev
20:33 ashtokol_ joined #fuel-dev
20:53 ashtokol_ joined #fuel-dev
21:21 ddmitriev joined #fuel-dev
21:29 youellet joined #fuel-dev
22:03 jobewan joined #fuel-dev
23:03 jobewan joined #fuel-dev
23:07 youellet_ joined #fuel-dev
23:49 ashtokol_ joined #fuel-dev

| Channels | #fuel-dev index | Today | | Search | Google Search | Plain-Text | summary