Thursday, 2014-01-23

*** Jianyong has quit IRC00:01
*** clev has quit IRC00:02
*** networkstatic is now known as networkstatic_zZ00:02
*** clev has joined #openstack-neutron00:03
*** clev has quit IRC00:04
*** networkstatic_zZ has quit IRC00:09
*** Makdaam_ has joined #openstack-neutron00:10
*** Makdaam has quit IRC00:12
openstackgerritIvar Lazzaro proposed a change to openstack/neutron: Embrane Tempest Compliance  https://review.openstack.org/6667100:19
*** banix has quit IRC00:21
*** sweston has joined #openstack-neutron00:23
*** rook has quit IRC00:29
*** matsuhashi has joined #openstack-neutron00:34
*** banix has joined #openstack-neutron00:59
*** clev has joined #openstack-neutron01:03
*** julim has quit IRC01:04
*** julim has joined #openstack-neutron01:07
*** aymenfrikha has quit IRC01:10
*** thuc has joined #openstack-neutron01:11
*** thuc_ has quit IRC01:13
*** thuc has quit IRC01:15
*** shivh has quit IRC01:21
*** thuc has joined #openstack-neutron01:29
*** thuc_ has joined #openstack-neutron01:29
openstackgerritZhiQiang Fan proposed a change to openstack/neutron: Enable hacking H233 rule  https://review.openstack.org/6853601:32
*** thuc_ has quit IRC01:33
*** thuc has quit IRC01:33
*** pcm_ has joined #openstack-neutron01:45
*** pcm_ has quit IRC01:45
*** pcm_ has joined #openstack-neutron01:46
*** ykaneko has joined #openstack-neutron01:49
*** sn6i23a has quit IRC01:52
*** alexpilotti has quit IRC01:54
*** SumitNaiksatam has quit IRC01:55
*** sweston has quit IRC01:59
*** aymenfrikha has joined #openstack-neutron02:08
*** aveiga has quit IRC02:15
*** vkozhukalov has joined #openstack-neutron02:23
*** ashaikh has joined #openstack-neutron02:30
*** julim has quit IRC02:35
*** mayu has joined #openstack-neutron02:35
*** aveiga has joined #openstack-neutron02:37
openstackgerritminsel proposed a change to openstack/neutron: Add XML capability to extra_DHCP_options extension  https://review.openstack.org/6854402:39
openstackgerritZhiQiang Fan proposed a change to openstack/python-neutronclient: Enable hacking H233 rule  https://review.openstack.org/6854502:39
mayuanteaya, are you  Anita Kuno ?02:40
*** ramishra has joined #openstack-neutron02:52
openstackgerritShiv Haris proposed a change to openstack/neutron: Fix db upgrade (migration) of brocadenetworks  https://review.openstack.org/6819902:55
*** aymenfrikha has quit IRC02:59
*** dguitarbite has joined #openstack-neutron02:59
openstackgerritSylvain Afchain proposed a change to openstack/neutron: Test patch, do not review  https://review.openstack.org/6848303:01
*** SumitNaiksatam has joined #openstack-neutron03:07
*** dguitarbite has quit IRC03:09
*** emagana has quit IRC03:10
*** nati_ueno has quit IRC03:10
*** ashaikh_ has joined #openstack-neutron03:19
*** ramishra has quit IRC03:19
*** ramishra has joined #openstack-neutron03:20
*** ashaikh has quit IRC03:20
*** ashaikh_ is now known as ashaikh03:20
*** terrylhowe has quit IRC03:22
*** ramishra has quit IRC03:24
*** mayu has quit IRC03:28
*** jdev789 is now known as jdev789_away03:34
openstackgerritwanghong proposed a change to openstack/neutron: ValueError should use '%' instead of ','  https://review.openstack.org/6649003:36
*** dguitarbite has joined #openstack-neutron03:41
*** ramishra has joined #openstack-neutron03:44
*** pcm_ has quit IRC03:45
*** aveiga has quit IRC03:48
openstackgerritXiang Hui proposed a change to openstack/neutron: Add options for some commands executing run_vsctl  https://review.openstack.org/5895403:51
*** iwamoto has joined #openstack-neutron03:58
*** mrsnivvel has joined #openstack-neutron04:04
*** thuc_ has joined #openstack-neutron04:06
*** yamahata has joined #openstack-neutron04:08
*** carl_baldwin has joined #openstack-neutron04:10
*** carl_baldwin has quit IRC04:11
*** carl_baldwin has joined #openstack-neutron04:17
*** carl_baldwin has quit IRC04:21
*** harlowja is now known as harlowja_away04:23
*** carl_baldwin has joined #openstack-neutron04:25
*** carl_baldwin has quit IRC04:29
*** markmcclain has joined #openstack-neutron04:32
*** morganfainberg is now known as morganfainberg|z04:35
openstackgerritstephen-ma proposed a change to openstack/neutron: L3 Agent restart causes network outage  https://review.openstack.org/3098804:39
*** markmcclain has quit IRC04:42
*** markwash_ has joined #openstack-neutron04:43
*** markmcclain has joined #openstack-neutron04:43
*** markwash has quit IRC04:43
*** markwash_ is now known as markwash04:43
*** thuc_ has quit IRC04:43
*** thuc has joined #openstack-neutron04:44
*** thuc has quit IRC04:48
*** yamahata has quit IRC04:57
*** nati_ueno has joined #openstack-neutron05:03
*** nati_uen_ has joined #openstack-neutron05:05
*** sbalukoff has quit IRC05:06
*** nati_ueno has quit IRC05:08
*** sbalukoff has joined #openstack-neutron05:10
*** harlowja_away is now known as harlowja05:10
*** alex_klimov has joined #openstack-neutron05:12
*** Jianyong has joined #openstack-neutron05:19
*** yamahata has joined #openstack-neutron05:20
*** dave_tucker is now known as dave_tucker_zzz05:26
*** itzikb has quit IRC05:28
*** Jianyong has quit IRC05:31
*** rohit404 has joined #openstack-neutron05:41
*** nati_ueno has joined #openstack-neutron05:53
*** nati_ueno has quit IRC05:53
*** nati_ueno has joined #openstack-neutron05:54
*** dave_tucker_zzz is now known as dave_tucker05:54
*** nati_uen_ has quit IRC05:56
*** markmcclain has quit IRC06:02
*** banix has quit IRC06:03
*** dave_tucker is now known as dave_tucker_zzz06:04
*** dave_tucker_zzz is now known as dave_tucker06:04
*** rohit404 has quit IRC06:10
*** irenab has joined #openstack-neutron06:19
openstackgerritJenkins proposed a change to openstack/neutron: Imported Translations from Transifex  https://review.openstack.org/6650106:21
*** afazekas has joined #openstack-neutron06:22
*** RobinWang has joined #openstack-neutron06:28
openstackgerritVincent Untz proposed a change to openstack/python-neutronclient: Fix listing security group rules  https://review.openstack.org/6835706:32
*** gdubreui has quit IRC06:40
*** bashok has joined #openstack-neutron06:40
*** harlowja is now known as harlowja_away06:45
*** irenab has quit IRC06:45
*** garyk has joined #openstack-neutron06:47
*** Jianyong has joined #openstack-neutron06:48
*** nati_uen_ has joined #openstack-neutron06:48
*** matsuhashi has quit IRC06:50
*** clev has quit IRC06:50
*** nati_ueno has quit IRC06:51
*** Jianyong has quit IRC06:53
*** matsuhashi has joined #openstack-neutron06:54
*** jlibosva has joined #openstack-neutron06:59
*** dave_tucker is now known as dave_tucker_zzz07:00
*** dave_tucker_zzz is now known as dave_tucker07:05
*** rohit404 has joined #openstack-neutron07:06
lifelesssalv-orlando: hey; so to make sure this hasn't dropped in the cracks; the ptmu thing; are you chasing your colleagues?07:09
*** jlibosva has quit IRC07:14
*** Jianyong has joined #openstack-neutron07:15
*** bjornar has joined #openstack-neutron07:20
*** emagana has joined #openstack-neutron07:22
openstackgerritenikanorov proposed a change to openstack/neutron: LBaaS: Fix discrepancy between status and admin_state_up  https://review.openstack.org/5503207:23
openstackgerritOleg Bondarev proposed a change to openstack/neutron: Reschedule router if new external gateway is on other network  https://review.openstack.org/5288407:24
*** amotoki has joined #openstack-neutron07:30
*** jlibosva has joined #openstack-neutron07:35
*** Jianyong has quit IRC07:37
*** rohit404 has quit IRC07:37
*** rohit404 has joined #openstack-neutron07:41
*** nati_ueno has joined #openstack-neutron07:42
*** krast has quit IRC07:44
*** emagana has quit IRC07:45
*** nati_uen_ has quit IRC07:46
*** yamahata has quit IRC07:46
*** djoreilly has joined #openstack-neutron07:49
*** Jianyong has joined #openstack-neutron07:51
*** ramishra_ has joined #openstack-neutron07:51
*** ramishra has quit IRC07:52
*** garyk has quit IRC07:54
*** RobinWang has quit IRC08:00
*** yamahata has joined #openstack-neutron08:02
*** luqas has joined #openstack-neutron08:15
*** morganfainberg|z has quit IRC08:22
*** Jianyong has quit IRC08:24
*** morganfainberg|z has joined #openstack-neutron08:25
*** morganfainberg|z is now known as morganfainberg08:25
*** Jianyong has joined #openstack-neutron08:26
*** krast has joined #openstack-neutron08:27
*** markwash has quit IRC08:29
*** markwash has joined #openstack-neutron08:30
*** markwash has quit IRC08:31
*** jistr has joined #openstack-neutron08:32
*** iwamoto has quit IRC08:36
*** Jianyong has quit IRC08:37
*** bjornar has quit IRC08:42
*** amuller has joined #openstack-neutron08:46
salv-orlandolifeless: I haven't yet heard of them, but pointed them to the discussion on the openstack mailing list08:48
*** rohit404 has quit IRC08:52
*** rohit404 has joined #openstack-neutron08:52
*** yamahata has quit IRC08:53
*** Jianyong has joined #openstack-neutron08:54
*** jlibosva has quit IRC09:00
*** Jianyong has quit IRC09:01
amullerajo: heya09:03
amullerajo: Invited you to hangout09:04
openstackgerritMehdi Abaakouk proposed a change to openstack/python-neutronclient: Add --segment support to net-create  https://review.openstack.org/6857609:05
*** ramishra_ has quit IRC09:06
*** ashaikh has quit IRC09:06
*** vkozhukalov has quit IRC09:06
*** Jianyong has joined #openstack-neutron09:07
*** jlibosva has joined #openstack-neutron09:09
*** irenab has joined #openstack-neutron09:10
*** jpich has joined #openstack-neutron09:10
*** yamahata has joined #openstack-neutron09:11
*** ramishra_ has joined #openstack-neutron09:11
*** ashaikh has joined #openstack-neutron09:11
*** vkozhukalov has joined #openstack-neutron09:11
*** ramishra_ has quit IRC09:13
*** ashaikh has quit IRC09:13
*** vkozhukalov has quit IRC09:13
*** vkozhukalov has joined #openstack-neutron09:14
*** safchain has joined #openstack-neutron09:14
*** ramishra has joined #openstack-neutron09:14
*** ashaikh has joined #openstack-neutron09:15
*** Jianyong has quit IRC09:16
*** ashaikh has quit IRC09:16
*** salv-orlando has quit IRC09:17
*** ramishra_ has joined #openstack-neutron09:21
*** ramishra has quit IRC09:22
openstackgerritRoman Podoliaka proposed a change to openstack/neutron: WIP: test if psycopg2 is related to races  https://review.openstack.org/6857709:25
*** garyk has joined #openstack-neutron09:26
*** ramishra_ has quit IRC09:30
*** beagles has quit IRC09:30
*** b3nt_pin has joined #openstack-neutron09:35
*** luqas has quit IRC09:38
*** bjornar has joined #openstack-neutron09:43
*** jlibosva has quit IRC09:46
*** amuller has quit IRC09:47
openstackgerritSylvain Afchain proposed a change to openstack/neutron: Add parameter and iptables rules to protect dnsmasq ports  https://review.openstack.org/6199409:53
*** ramishra has joined #openstack-neutron10:01
*** amuller has joined #openstack-neutron10:02
*** yfried has quit IRC10:03
*** yfried has joined #openstack-neutron10:05
*** jlibosva has joined #openstack-neutron10:05
*** dosaboy has quit IRC10:07
*** luqas has joined #openstack-neutron10:09
*** xianghui has joined #openstack-neutron10:12
*** ColinTaylor has quit IRC10:15
*** ykaneko has quit IRC10:15
*** yfried has quit IRC10:17
*** ramishra has quit IRC10:20
*** ramishra has joined #openstack-neutron10:21
*** yfried has joined #openstack-neutron10:21
*** xianghui has quit IRC10:23
*** jp_at_hp has joined #openstack-neutron10:38
*** ramishra_ has joined #openstack-neutron10:41
*** ramishra has quit IRC10:45
*** bjornar has quit IRC10:50
*** afazekas has joined #openstack-neutron10:50
openstackgerritMarcos Fermín Lobo proposed a change to openstack/python-neutronclient: session_persistence: invalid str2dict value: u''  https://review.openstack.org/6860310:54
openstackgerritSylvain Afchain proposed a change to openstack/neutron: Introduces a keepalived manager for HA  https://review.openstack.org/6814211:01
*** matsuhashi has quit IRC11:05
*** salv-orlando has joined #openstack-neutron11:06
openstackgerritEvgeny Fedoruk proposed a change to openstack/neutron: Multiple service providers with same driver  https://review.openstack.org/6413911:06
openstackgerritSylvain Afchain proposed a change to openstack/neutron: Introduces a keepalived manager for HA  https://review.openstack.org/6814211:07
*** matsuhashi has joined #openstack-neutron11:14
openstackgerritSylvain Afchain proposed a change to openstack/neutron: Add parameter and iptables rules to protect dnsmasq ports  https://review.openstack.org/6199411:15
*** jprovazn_ has joined #openstack-neutron11:20
*** ramishra has joined #openstack-neutron11:24
*** ramishra_ has quit IRC11:24
openstackgerritVincent Untz proposed a change to openstack/neutron: Fix alembic migration for postgresql  https://review.openstack.org/6861111:27
*** rossella_s has joined #openstack-neutron11:28
*** amuller_ has joined #openstack-neutron11:36
*** amuller has quit IRC11:36
*** salv-orlando has quit IRC11:41
*** pcm_ has joined #openstack-neutron11:43
*** pcm_ has quit IRC11:45
*** pcm_ has joined #openstack-neutron11:46
*** garyk has quit IRC11:48
*** garyk has joined #openstack-neutron11:49
*** ramishra_ has joined #openstack-neutron11:50
*** ramishra_ has quit IRC11:52
*** ramishra has quit IRC11:53
*** ramishra_ has joined #openstack-neutron11:53
*** salv-orlando has joined #openstack-neutron11:55
*** salv-orlando has quit IRC11:56
*** ramishra_ has quit IRC11:59
*** djoreilly has quit IRC12:00
*** amuller__ has joined #openstack-neutron12:01
*** amuller_ has quit IRC12:02
*** irenab_ has joined #openstack-neutron12:05
*** irenab has quit IRC12:05
*** ramishra has joined #openstack-neutron12:05
*** dave_tucker is now known as dave_tucker_zzz12:05
*** irenab_ is now known as irenab12:05
*** ramishra has quit IRC12:11
*** ramishra has joined #openstack-neutron12:11
*** salv-orlando has joined #openstack-neutron12:13
*** ramishra has quit IRC12:15
*** ramishra has joined #openstack-neutron12:16
*** xianghui has joined #openstack-neutron12:22
anteayamayu I am12:22
*** luqas has quit IRC12:23
*** alexpilotti has joined #openstack-neutron12:25
*** salv-orlando has quit IRC12:25
*** b3nt_pin has quit IRC12:27
*** b3nt_pin has joined #openstack-neutron12:27
*** b3nt_pin is now known as beagles12:28
*** alexpilotti has quit IRC12:30
*** luqas has joined #openstack-neutron12:33
*** ramishra has quit IRC12:34
*** matsuhashi has quit IRC12:39
*** dguitarbite has quit IRC12:48
*** ramishra has joined #openstack-neutron12:48
*** luqas has quit IRC12:49
*** ramishra has quit IRC12:58
*** ramishra has joined #openstack-neutron12:58
jlibosvaanteaya: Hello, I can see isolated jobs passing. May I ask what was the culprit?12:58
*** alexpilotti has joined #openstack-neutron12:59
anteayaI have no idea12:59
anteayajlibosva: do you have a link?12:59
jlibosvaanteaya: https://jenkins03.openstack.org/job/check-tempest-dsvm-neutron-isolated/12:59
anteayainteresting I have not see this gui view before, how did you get to it?13:01
*** heyongli has joined #openstack-neutron13:03
anteayawell salv-orlando was going to roll back all the tempest changes we made last week, perhaps that happened13:04
* anteaya goes to check the tempest git logs13:05
*** ramishra_ has joined #openstack-neutron13:05
*** ramishra has quit IRC13:05
jlibosvaanteaya: that's Jenkins. On Zuul page I chose the isolated job that led me to the specific build # ( https://jenkins03.openstack.org/job/check-tempest-dsvm-neutron-isolated/182/) from project ..neutron-isolated. Upper left corner has some site navigation or there is a link "Back to Project".13:05
*** ajo has quit IRC13:05
anteayanope, http://git.openstack.org/cgit/openstack/tempest/log/13:06
jlibosvaanteaya: I saw pushing patches so I wanted to check the gate status to know whether I should continue digging into test_minimum_basic or not13:06
anteayathank you13:06
anteayaI have no data on why isolated jobs are passing13:06
jlibosvaanteaya: nice dwarfs fixed it during the night :) thank you for info13:08
anteayathat's about all I can come up with, I see nothing in the channel logs or tempest logs that anyone had a eureka moment13:08
anteayait was such a problem I would expect if someone where to have fixed it, they would claim responsibility13:09
anteayadevstack has nothing in the last 24 hours either: http://git.openstack.org/cgit/openstack-dev/devstack/log/13:10
*** polfilm has joined #openstack-neutron13:11
anteayathe only thing I can come up with is that there was a big rearrangement with zuul yesterday, in as far as now all patches in the check queue have jobs running on them13:11
anteayaso check is not starved for vms13:11
anteayaif that is the reason isolated jobs are passing again, I would be very curious as to the relationship between the jobs and the situation in check13:12
anteayaand no one has been approving anything so we have no data about how these jobs are performing in the gate queue13:12
polfilmHi. I hope you don't mind this one time assignment post: $50 for help with default install setup : https://www.freelancer.com/projects/Software-Testing-software-development/Neutron-Networking-OpenStack-Centos.html13:13
anteayaso last failed build for check-tempest-dsvm-neutron-isolated was 6 hours and 53 minutes ago13:17
*** ajo has joined #openstack-neutron13:17
anteayaand for check-tempest-dsvm-neutron-isolated was 6 hours and 46 minutes ago13:18
*** ramishra_ has quit IRC13:18
anteayawhat happened at about 06:00 UTC?13:18
*** ramishra has joined #openstack-neutron13:18
*** aveiga has joined #openstack-neutron13:19
anteayathe last change to neutron was 5 days ago: http://git.openstack.org/cgit/openstack/neutron/log/13:20
*** dosaboy has joined #openstack-neutron13:24
anteayanow they are failing13:25
anteaya:(13:25
anteayaso in answer to your question jlibosva I think yes13:25
anteayaI think you need to dig into it13:26
anteayaI think the failure rate for isolated jobs was about 70%13:27
anteayaso some of those passing tests got through the 30% gap13:27
jlibosvaanteaya: ok, thanks13:27
anteayathanks for showing me how to view all the jobs on the same page13:28
anteayahelpful and I didn't know that before13:28
anteayaso if I remember correctly the last place you were was figuring out why there was a difference running the command as sudo and as stack13:29
anteayayes?13:29
anteayaobondarev: hello13:29
anteayaI don't think rechecking patches will get you much yet13:29
*** polfilm has quit IRC13:30
*** ramishra has quit IRC13:30
*** ramishra has joined #openstack-neutron13:31
enikanorov_jlibosva: do you see isolated passing on patches on which they have failed previously?13:32
jlibosvaenikanorov_: didn't check that, I just checked the jenkins without looking to related patches13:33
anteayahere is one example of a patch that has isolated passing where they failed previously, mind you they are different patchsets of the same patch series: https://review.openstack.org/#/c/61994/13:36
enikanorov_ok, lets make a recheck to this patch13:37
openstackgerritAkihiro Motoki proposed a change to openstack/neutron: Test patch (do not review)  https://review.openstack.org/6807913:38
obondarevanteaya: Hi. I wasn't sure why isolated jobs started to pass with no errors to, so tried to take a chance to have +1 on my patches (given that zuul check queue is not too long)13:38
anteayaenikanorov_: good idea13:38
enikanorov_i this isolated are not passing more than usual13:38
ajowhat are the isolated jobs exactly?13:38
anteayaobondarev: fair enough13:39
enikanorov_also, i wonder if we can find out which jenkins slave run failed and passed jobs13:39
anteayaajo: go to this page: http://status.openstack.org/zuul/13:39
anteayaajo in the check queue, the one on the left find a neutron patch13:39
anteayaclick on the blue bar which expands the running tests for that patch13:40
ajo*-isolated13:40
*** thomasem has joined #openstack-neutron13:40
anteayayes13:40
ajowhat's the difference with the other tests?13:40
*** dosaboy has quit IRC13:41
anteayathe isolated tests are specific for neutron13:41
anteayaperhaps enikanorov_ or obondarev knows more specific information than myself13:41
ajooh, ok, so they test neutron by itself without other openstack components?13:42
*** dosaboy has joined #openstack-neutron13:42
enikanorov_ajo: no13:42
*** pcarver has joined #openstack-neutron13:42
enikanorov_ajo: isolated is a job which runs tempest tests in parallel under several tenants13:42
ajooohh13:42
enikanorov_instead of doing it consequently13:42
ajook, got it13:42
anteayaenikanorov_: if you go to the very top of the console output, one of the top lines gives the information about which slave the test ran on13:42
enikanorov_anteaya: thanks13:42
ajoso every test is separated in a different tenant13:43
ajoto avoid quota/other conflicts13:43
anteayaso this slave Building remotely on devstack-precise-hpcloud-az2-118601013:43
ajothanks enikanorov , anteaya13:43
obondarevenikanorov: not exactly: for parallel there is a separate job - see https://review.openstack.org/#/c/64185/13:43
ajoobondarev, so .. then?13:44
anteayawas where check-tempest-dsvm-neutron-isolated ran for https://review.openstack.org/#/c/61994/ patchset 19 success run13:44
obondarevso isolated means that each test creates a new tenant13:44
ajoaha13:44
ajook, I suppose is the path for good parallel testing13:44
ajoif isolated runs smoothly, then the tests can be run in parallel safely13:45
ajo+/-13:45
obondarevyes13:45
obondarevnot exactly however13:46
obondarevparallel tests had lots of issues13:46
ajoyes, I suppose that external networks are complicated to handle even for isolated, etc..13:46
ajoand sure there are more issues..13:47
obondarevfull tenant isolation helps but doesn't solves all issues13:47
*** tongli has joined #openstack-neutron13:50
openstackgerritAkihiro Motoki proposed a change to openstack/neutron: Test patch (do not review)  https://review.openstack.org/6807913:53
openstackgerritSylvain Afchain proposed a change to openstack/neutron: Add L3 VRRP HA base classes  https://review.openstack.org/6455313:53
*** Jianyong has joined #openstack-neutron13:53
*** ramishra has quit IRC13:54
*** ramishra has joined #openstack-neutron13:54
*** luqas has joined #openstack-neutron13:57
*** jecarey has quit IRC13:57
*** thuc has joined #openstack-neutron13:57
*** thuc_ has joined #openstack-neutron13:58
*** yamahata has quit IRC13:58
openstackgerritAkihiro Motoki proposed a change to openstack/neutron: Test patch (do not review)  https://review.openstack.org/6807913:59
*** aymenfrikha has joined #openstack-neutron14:00
*** thuc has quit IRC14:02
*** markmcclain has joined #openstack-neutron14:03
*** yfujioka has joined #openstack-neutron14:03
*** WackoRobie has joined #openstack-neutron14:03
*** ramishra has quit IRC14:04
*** jroovers has joined #openstack-neutron14:05
*** heyongli has quit IRC14:10
*** ramishra has joined #openstack-neutron14:12
*** pcarver has quit IRC14:13
*** ramishra_ has joined #openstack-neutron14:14
*** ramishra has quit IRC14:14
*** yamahata has joined #openstack-neutron14:14
*** ramishra_ has quit IRC14:15
*** ramishra has joined #openstack-neutron14:16
*** baoli_ has joined #openstack-neutron14:18
baoli_Hi Irena14:18
*** changbl has quit IRC14:19
irenabbaoli_: hi14:19
irenabbaoli: I have about 10 mins14:20
baoli_ok14:20
baoli_I think that we should explore the idea of associating a flavor with the physical net14:20
irenabbaoli: I am starting  with general part of enabling neutron ML2 with port binding:profile support and vnic_type request14:20
baoli_irenab, that's great14:21
*** ramishra_ has joined #openstack-neutron14:21
baoli_Seems like that you have some scheduler questions?14:21
irenabbaoli: I think you idea is fine, as long as nova supports14:21
irenaband we definitly can start with condig option in neutron, to make it less code changes14:21
baoli_So if we do the physical net to flavor association, we may not need the flavor in the --nic14:22
irenabI think we can also go with --nic port option without nova api for the first version14:22
baoli_What do you mean by "without nova api"?14:23
irenabI do not belive as a team we can make all parts coded with uni tests and pushed for review in coming 4 weeks14:23
baoli_Irenab, I understand that14:23
irenabso I mean we can make the api as follow up change14:23
baoli_Can you be specific about the api?14:24
*** ramishra_ has quit IRC14:24
*** ramishra_ has joined #openstack-neutron14:24
*** julim has joined #openstack-neutron14:24
irenabI just suggested to skip the nova pai extension for comming weeks effort and concentrate on more internal parts14:24
*** ramishra has quit IRC14:24
irenab^api14:25
baoli_You mean to say the pci flaovr api discussion? I'm with you on that14:25
irenabbaoli:seems that patch reviews in nova are very slow14:25
baoli_irenab, that's expected. I have seen a review with a couple lines of code going on for several weeks14:26
irenabso even if we reach the agreement with all guys who come and go into dicsussion, its too many changes to squize14:26
*** dims has quit IRC14:26
baoli_Exactly. So we should go ahead with our sriov development and let the nova guys figure out flavor versue aggregate, etc14:27
irenabbaoli_:agree14:27
baoli_sounds good. So on Monday, let's nail down what to do next14:28
baoli_and start coding14:28
*** dims has joined #openstack-neutron14:28
irenabbaoli: so if you are OK to start with setting the net pci flavor via config file and port profile via binding extension, we can start with nova boot --nic port_id option14:28
*** dave_tucker_zzz is now known as dave_tucker14:29
irenabbaoli: Yes!14:29
baoli_that's a good idea.14:29
baoli_Let's talk about that on Monday14:29
baoli_I guess that you have to go now14:29
irenabbaoli: yes, have to run now, will try to go over email later today. See you on Monday, have a nice weekend14:29
baoli_have a good weekend14:30
*** thuc_ has quit IRC14:30
*** garyk has quit IRC14:30
*** thuc has joined #openstack-neutron14:30
*** ramishra has joined #openstack-neutron14:33
*** jecarey has joined #openstack-neutron14:33
anteayaenikanorov_: tests returned in less than an hour (yay) but failed on isolated (boo)14:34
*** thuc has quit IRC14:34
anteayawhich basically leads me to conclude that the passed tests were the anomoly14:34
anteayathat and the patch itself doesnt' seem to be contributing to the failures14:35
*** ramishra_ has quit IRC14:37
*** ramishra has quit IRC14:38
*** irenab has quit IRC14:39
*** aymenfrikha has quit IRC14:45
*** peristeri has joined #openstack-neutron14:48
*** bvandenh has joined #openstack-neutron14:54
*** ramishra has joined #openstack-neutron14:56
*** banix has joined #openstack-neutron14:57
*** banix has quit IRC14:57
*** clev has joined #openstack-neutron15:00
otherwiseguymarkmcclain: did we ever figure out what was going to be done re: psutil?15:00
markmcclainotherwiseguy: I've beeb working on it in spurts15:01
markmcclainwith the gate problems it moved to backburner15:01
markmcclainif you got cycles to work on it that would be great15:01
otherwiseguydidn't know if we were trying to change versions or remove the dependencies, or what.15:01
*** ramishra has quit IRC15:01
ajootherwiseguy, what happens with psutil? I had problems today trying to run tests because of it15:02
*** alex_klimov has quit IRC15:03
markmcclainif we use the work around from havana.. we can drop psutil15:03
*** julim has quit IRC15:03
otherwiseguythe version in the requirements file isn't actually hosted on pypi, but newer ones are. i have a workaround in havana that doesn't require psutil (though is certainly an uglier solution, but it is an edge case so doesn't really matter much).15:04
otherwiseguyajo: ^15:04
*** banix has joined #openstack-neutron15:04
otherwiseguymarkmcclain: I can go that way if people want. Last I'd heard someone was talking about just using the newer version or something so wasn't sure if anything had been decided.15:05
ajoohpp15:05
markmcclainseems like the pain of requiring a new version might be lower long term15:05
*** WackoRobie has quit IRC15:05
markmcclainI haven't surveyed to see how many other projects we'd have to update too15:06
otherwiseguyheat cntools was the only place I noticed it.15:06
markmcclainmoving to a version hosted on PyPI would be good long term anyway15:06
otherwiseguycfntools15:06
markmcclainok15:06
otherwiseguyI could also see if the author would host the older versions on pypi. :p15:06
*** luqas has quit IRC15:07
markmcclainthat would certainly fix things quickly15:07
*** julim has joined #openstack-neutron15:07
markmcclainI've to step out for a bit, but I'm comfortable with any of these options15:08
*** mfink has joined #openstack-neutron15:09
*** WackoRobie has joined #openstack-neutron15:10
*** luqas has joined #openstack-neutron15:11
*** changbl has joined #openstack-neutron15:16
ajohttps://review.openstack.org/#/c/32235/  : I'm backporting those oslo fixes to havana and grizzly, where possible15:17
ajo(rpc fixes related to bug 1189711 Should RPC consume_in_thread() be more fault tolerant )15:18
ajoto avoid the consumer from dying on underlying exceptions15:18
*** clev has quit IRC15:18
*** garyk has joined #openstack-neutron15:19
*** ramishra has joined #openstack-neutron15:22
*** afazekas has quit IRC15:23
*** Jianyong has quit IRC15:23
ajohmmm15:25
ajobut we didn't had tests/rpc/test_*.py in havana15:25
*** jgrimm has joined #openstack-neutron15:26
*** xianghui has quit IRC15:26
ajogaryk, ping15:27
ajoI was checking, you backported    https://review.openstack.org/#/c/32235/ (from oslo) in  https://github.com/openstack/neutron/commit/4dbb55cbb0020702a3ea2ec89ddf61319d00f858  together with a few extra updates15:28
openstackgerritmayu proposed a change to openstack/neutron: Move RemoteRestError from clients.py to exceptions.py  https://review.openstack.org/6814815:28
ajowe don't bring the oslo tests to neutron, right?15:29
*** mrsnivvel has quit IRC15:29
*** ramishra has quit IRC15:30
*** ramishra has joined #openstack-neutron15:31
*** aymenfrikha has joined #openstack-neutron15:32
garykajo: correct - oslo tests should not be included15:32
ajounderstood :)15:33
ajothanks :)15:33
*** aymenfrikha has quit IRC15:34
*** changbl has quit IRC15:36
*** ramishra has quit IRC15:38
*** markmcclain has quit IRC15:45
*** MM_at_HP has quit IRC15:50
*** djoreilly has joined #openstack-neutron15:52
*** ramishra has joined #openstack-neutron15:53
*** ashaikh has joined #openstack-neutron15:55
ajogaryk: https://review.openstack.org/#/c/68672/15:55
ajoanteaya, enikanorov  , obondarev,  not sure if given the current status of the gate I may refrain from submitting backport patches at this moment15:56
garykajo: your cherry pick is problematic - please see the comments15:57
ajosure, thanks, let me check15:57
*** ramishra has quit IRC15:57
*** banix has quit IRC15:57
ajooups, thanks15:59
ajoIs there a way to cherry pick from oslo directly? may be just patch export and patch -p115:59
garyki do the following:16:00
garykcheckout the lastest oslo config16:00
garykcheckout the neutron brach16:00
garykbranch16:00
garykthere is a script in oslo that you need to run16:00
garykupdate.sh with the neutron conf file (or the packages that you want to import)16:01
ajonice, it does let you cherry pick just a commit from oslo?16:01
*** banix has joined #openstack-neutron16:01
*** ramishra has joined #openstack-neutron16:01
garykno, it copies the latest code from the master16:01
garykor the branch that you have checked out16:01
garykthat is the primitive way that i go about it - maybe others have a better way. i sometimes take the simple and laborious way - but it works :)16:02
ajomay be the best way would be... backport into oslo16:02
ajooslo/havana16:02
ajothen pick it back to neutron/havana16:03
garykthe upside of the oslo script is that it will also pull in dependencies16:03
garyknot sure i follow - same process is done for the stable16:03
ajoI mean16:04
ajo1) backport the fix from oslo/master to oslo/havana16:04
ajo2) then use the update.sh script from oslo/havana to neutron/havana16:04
*** alagalah__ has quit IRC16:04
ajo1 is done already, btw :)16:05
ajothanks16:05
ajoI will try the update.sh script garyk16:05
ajovery helpful16:05
*** ramishra has quit IRC16:06
garykread the code in update.py and it will help explain better than i do.16:07
*** arash has joined #openstack-neutron16:10
*** ramishra has joined #openstack-neutron16:11
*** emagana has joined #openstack-neutron16:12
*** aymenfrikha has joined #openstack-neutron16:12
anteayaajo you can submit anything to the check queue, but if isolated jobs fail please don't keep rechecking them16:13
*** terrylhowe has joined #openstack-neutron16:13
anteayaif backport patches run isolated jobs, please don't approve them since they will likely just fail in the gate until isolated jobs are fixed16:14
anteayaalso sdague has the issue been fixed with stable?16:17
anteayaI think it was a pip issue, was it not?16:17
*** ramishra has quit IRC16:18
*** ramishra_ has joined #openstack-neutron16:18
arashHello16:18
arashany body knows how we can get some values from config file (local.conf)??16:19
arashAre those values stores somewhere in neutron database?16:20
*** thuc has joined #openstack-neutron16:22
*** SumitNaiksatam has quit IRC16:22
*** bvandenh has quit IRC16:23
*** ramishra_ has quit IRC16:27
*** nati_ueno has quit IRC16:28
*** morganfainberg is now known as morganfainberg|z16:31
*** carl_baldwin has joined #openstack-neutron16:32
*** fredwen has joined #openstack-neutron16:41
*** rohit404_ has joined #openstack-neutron16:43
*** morganfainberg|z is now known as morganfainberg16:43
*** rohit404 has quit IRC16:44
*** rohit404_ is now known as rohit40416:44
*** bashok has quit IRC16:44
*** aymenfrikha has quit IRC16:45
*** amotoki is now known as amotoki_zzz16:46
*** aymenfrikha has joined #openstack-neutron16:47
*** fredwen has quit IRC16:48
*** aymenfrikha has quit IRC16:50
*** aymenfrikha has joined #openstack-neutron16:50
*** safchain has quit IRC16:50
*** polfilm has joined #openstack-neutron16:51
*** jeblair has joined #openstack-neutron16:52
openstackgerritMarios Andreou proposed a change to openstack/neutron: Make allocation_pools attribute of subnet updateable by PUT  https://review.openstack.org/6204216:53
*** ChanServ sets mode: +o jeblair16:53
*** mlavalle has joined #openstack-neutron16:53
*** jeblair sets mode: +o anteaya16:53
*** jeblair sets mode: +b *!*polfilm@*.4-2.cable.virginm.net16:54
*** polfilm was kicked by jeblair (polfilm)16:54
anteayathank you16:54
marioscarl_baldwin: hi carl - thanks for your review comments on ^^^ i _think_ I got all of them in. I'd be grateful when you get some time. thx16:54
carl_baldwinmarios: I will have a look today.16:55
marioscarl_baldwin: cheers16:55
carl_baldwinNice quick turn-around on those changes.16:56
*** ashaikh has quit IRC16:57
marioscarl_baldwin: yeah... i *thought* they'd be simple but ended up spending most of the afternoon (tbh i was just grateful to get a review ;) ).never mind, pushed some stuff to tomorrow.16:57
*** jlibosva has quit IRC16:57
*** jlibosva has joined #openstack-neutron16:58
*** yfujioka has quit IRC16:58
*** SumitNaiksatam has joined #openstack-neutron16:59
sdagueanteaya: stable/havana I believe is working now17:00
anteayaawesome thank you17:00
anteayaajo: so it looks like sending backports to check, once per patchset is okay atm17:00
anteayaalso I'm ops for reasons of getting rid of spammers17:01
anteayaanyone else who is in a different timezone and is in irc a lot and can help keeping the channel clear, please volunteer for ops and we will get a sense from the group who would be good ops people17:02
ajoanteaya,  thank you very much17:03
anteayanp17:04
mariosanteaya: i am in GMT+1 fwiw (now is 1900 here)17:04
*** amuller has joined #openstack-neutron17:04
*** aymenfrikha has quit IRC17:04
marioserr GMT+117:04
mariosdammit! GMT+217:04
*** dave_tucker is now known as dave_tucker_zzz17:05
*** amuller__ has quit IRC17:05
anteayamarios: do you like to be on irc a lot and are you comfortable with helping keep the channel clear?17:05
* marios stops fighting with keyboard17:05
mariosanteaya: well, my client is always connected (irssi on cloud server).17:05
anteayathat is a good start17:05
anteayalet me talk with markmcclain to get his thoughts first before I start green @'ing others17:06
mariosanteaya: i'll pvt you my email address17:06
anteayathanks for being willing to help17:06
mariosanteaya: sure, just mentioning it, fyi17:06
anteayasure17:06
*** marun has joined #openstack-neutron17:06
anteayayeah, thanks17:06
anteayahey marun17:06
marunanteaya: hi17:06
*** yamahata has quit IRC17:07
*** marios is now known as marios|gone17:07
*** markwash has joined #openstack-neutron17:11
*** tong_ has joined #openstack-neutron17:12
*** tongli has quit IRC17:14
*** MM_at_HP has joined #openstack-neutron17:24
*** sbalukoff has quit IRC17:24
*** dave_tucker_zzz is now known as dave_tucker17:24
*** jistr has quit IRC17:28
*** afazekas has joined #openstack-neutron17:30
*** baoli_ has quit IRC17:30
*** afazekas has quit IRC17:36
*** changbl has joined #openstack-neutron17:38
*** jroovers has quit IRC17:39
*** csd_ has joined #openstack-neutron17:40
*** rohit404 has quit IRC17:40
*** csd_ has quit IRC17:40
*** csd_ has joined #openstack-neutron17:41
*** rohit404 has joined #openstack-neutron17:43
*** jpich has quit IRC17:43
*** ashaikh has joined #openstack-neutron17:44
*** ygbo has quit IRC17:44
*** markmcclain has joined #openstack-neutron17:51
*** pcarver has joined #openstack-neutron17:52
*** yamahata has joined #openstack-neutron17:53
*** aymenfrikha has joined #openstack-neutron17:56
*** csd_ is now known as csd17:58
*** tong_ has quit IRC17:58
*** jdev789_away has quit IRC17:58
*** markmcclain has quit IRC17:58
*** nati_ueno has joined #openstack-neutron18:00
*** SumitNaiksatam has quit IRC18:02
*** aymenfrikha has quit IRC18:02
*** tongli has joined #openstack-neutron18:03
*** harlowja_away is now known as harlowja18:10
*** SumitNaiksatam has joined #openstack-neutron18:12
*** luqas has quit IRC18:12
openstackgerritDane LeBlanc proposed a change to openstack/neutron: Test patch, do not review  https://review.openstack.org/6871518:12
*** sbalukoff has joined #openstack-neutron18:17
*** rohit404 has quit IRC18:17
*** sputnik13net has joined #openstack-neutron18:18
*** rohit404 has joined #openstack-neutron18:19
enikanorov_jaypipes: Hi18:21
jaypipesenikanorov_: hi!18:21
enikanorov_i;d like to discuss https://review.openstack.org/6413918:22
enikanorov_the patch allows cloud admin to implement what is called 'service flavor' with configuration only18:23
enikanorov_and, for instance, setup billing that will depend on such flavor18:24
jaypipesenikanorov_: sure, happy to discuss.18:25
enikanorov_jaypipes: i saw you suggestion to implement it with different classes, but it's really not what we are trying to achieve18:26
*** jroovers has joined #openstack-neutron18:26
jaypipesenikanorov_: sorry, I don't like the implementation. It's not necessary, IMO. Billing concerns should not be influencing how a driver is constructed or loaded.18:28
enikanorov_it's not about billing only, it's about service flavor18:28
enikanorov_we have no way of providing different parameters to the driver at this moment18:29
jaypipesenikanorov_: in the patch, there's nothing that cannot be implemented with a single config section for [radware] and possibly specialized driver subclasses.18:29
enikanorov_that that is a good way of doing this, considering how we work with providers at all18:29
enikanorov_jaypipes: subclasses are done by vendor18:29
enikanorov_where configuration is done by cloud operator18:29
jaypipesenikanorov_: the idea of "service flavor" does not make much sense to me.18:29
*** armax has joined #openstack-neutron18:30
enikanorov_jaypipes: but that's a common and desired feature. you know, nova has it in the core api18:30
*** jlibosva has quit IRC18:30
jaypipesenikanorov_: no, it does not have this concept.18:30
jaypipesenikanorov_: I believe there is substantial confusion regarding this.18:30
enikanorov_it has, 'flavor' is exactly what we are trying to introduce here18:30
*** jroovers has quit IRC18:30
enikanorov_flavor is a way to request capabilities  of the resource18:31
enikanorov_in nova it is amount of disk/mem/cpu18:31
jaypipesenikanorov: a flavor in nova is something that is selected by the end user. there is not anything in the nova hypervisor driver configuration that changes based on the flavor a user selects.18:31
enikanorov_here it could be other capabilities18:32
enikanorov_jaypipes: that's because flavor in nova is stored in the db18:32
jaypipesenikanorov: capabilities are one thing. changing the way the driver configuration is handled is another.18:32
enikanorov_jaypipes: we can't store our 'flavors' in the db18:32
jaypipesenikanorov: where does the end-user specify this "service flavor" that you speak of?18:33
enikanorov_anyway, subclasses here aren't solving anything, because the problem is different18:33
enikanorov_jaypipes: it specifies provider, in fact18:33
jaypipesenikanorov: please explain how the problem is different...18:33
enikanorov_ok, I'll give a user story18:34
*** amuller has quit IRC18:34
jaypipeskk18:34
*** alex_klimov has joined #openstack-neutron18:34
enikanorov_I'm a cloud amid, I'm going to provide users with 'fast balancer' and 'slow balancer', for which i have 2 VMs, with more and less resources18:34
*** jroovers has joined #openstack-neutron18:34
enikanorov_both are vendorX-managed VMs18:35
enikanorov_how do I allow user to chose those providers?18:35
enikanorov_the driver is exactly the same in both cases.18:35
enikanorov_and, as a cloud admin, i'm not going to do any coding work18:36
jaypipesenikanorov: one sec, lemme look at LBaaS API.18:36
enikanorov_but I may decide to add 'medium balancer' (or whatever), changing the capabilities of backed appliances18:36
*** SumitNaiksatam_ has joined #openstack-neutron18:37
jaypipesenikanorov: make one LB pool for high I/O balancer VMs, and another for low I/O...18:37
jaypipesenikanorov: I'm not actually a fan of having LB inside a VM owned by the driver partly because of these reasons...18:38
enikanorov_pools are created by end users18:39
enikanorov_where the providers are offered by the cloud18:39
enikanorov_"having LB inside a VM owned by the driver" - but  that's already an existing solution18:40
enikanorov_and not only by 1 vendor18:40
*** SumitNaiksatam has quit IRC18:40
*** SumitNaiksatam_ is now known as SumitNaiksatam18:40
jaypipesenikanorov: in any case, the driver doesn't change anything... it should just be supplied some extra information in the call to POST /members18:40
*** MM_at_HP has quit IRC18:40
enikanorov_yes, and that is what is implemented18:41
enikanorov_the same driver has different configs and uses it in it's communication with backend18:41
jaypipesenikanorov: so why is there a need to change the way that driver configuration is loaded? If extra information is provided in the POST /members request, I don't see any point in the proposed patcvh.18:41
jaypipesenikanorov: but it's not different configs. It's the same configuration options.18:42
enikanorov_oh, sorry, misread your phrase18:42
*** terrylhowe has quit IRC18:43
enikanorov_no, that's not how it should work. end users doesn't provide anything additional in POST /members18:43
enikanorov_because it is unaware of how it works on the backend side18:43
jaypipes*sigh*... this is so convoluted.18:44
enikanorov_as user you want to work with fast device18:45
enikanorov_you create your LB choosing vendorX-fast provider18:45
enikanorov_or if you're poor, you choose vendorX-cheap provider18:45
enikanorov_you scripts remain the same18:45
jaypipesbut you just said the end user doesn't know that vendorX-fast-provider even exists.18:45
enikanorov_it knows, because it can list providers18:45
jaypipesyou just said above " because it is unaware of how it works on the backend side"18:46
jaypipes?18:46
jaypipesI see nothing in https://wiki.openstack.org/wiki/Neutron/LBaaS/API#Bulk_Create_Members about "listing providers"18:46
jaypipessorry... https://wiki.openstack.org/wiki/Neutron/LBaaS/API18:47
enikanorov_it is unaware, but it can chose providers. unaware means that user is isolated from backend implementation details18:47
jaypipes*who* can choose providers?18:47
enikanorov_the user18:47
jaypipesthe caller of POST /members?18:47
enikanorov_POST /pools18:47
jaypipeshow does that user know the "providers"?18:47
enikanorov_(pool is the first objectin lbaas workflow)18:47
enikanorov_neutron service-provider-list18:48
jaypipesI see nothing in https://wiki.openstack.org/wiki/Neutron/LBaaS/API#Create_a_Pool about a "provider" attribute of a pool.18:48
jaypipesso how does a user know which providers are available?18:48
enikanorov_by running 'neutron service-provider-list ' command18:49
enikanorov_the wiki page is pretty outdated18:49
enikanorov_please refer to API specs18:49
jaypipeswhere is API specs?18:50
enikanorov_let me find the link18:50
enikanorov_ok, that one is closer to current state of things, but still not up to date: http://docs.openstack.org/api/openstack-network/2.0/content/lbaas_ext_ops_pool_create.html18:53
enikanorov_need to file a doc bug on it. so the provider is one of attributes a user can pass to POST /pools18:55
jaypipesand how would a user know what value to put in that field?18:55
enikanorov_neutron service-provider-list18:56
jaypipeswe're going in circles.18:56
enikanorov_that command lists available providers for all services18:57
jaypipeshow is that useful to a user?18:57
jaypipesI just don't get it.18:58
jaypipeswhy don't you just implement LB types as a real resource, instead of this hack in the driver configuration? If this really is something that end users are begging for, then we should implement it correctly, with a real LB type that is stored and standardized in the backend database.18:59
jaypipesthis patch is just a giant hack to get around doing this properly, IMO.18:59
enikanorov_not quite sure I get your idea18:59
enikanorov_LB types as a real resource - what does it mean, from API perspective?19:00
jaypipesPOST /lb-types {'type': 'high-end'} ... POST /pools {'pool': {... 'lb_type': 'high-end'}}19:00
jaypipesthen pass lb_type to the main driver interface method for creating a LB pool resource19:01
*** jprovazn_ has quit IRC19:03
enikanorov_yes, that was discussed when we designed provider framework19:03
jaypipesenikanorov: sorry, I was under the impression that pool was a driver-controlled thing... but it's not, sorry.19:03
jaypipesit's just an end-user thing...19:03
enikanorov_but that is not going to work well19:03
jaypipeswould be better than a hack.19:03
enikanorov_nope, the proposed patch is not a hack. it's a configuration path19:04
enikanorov_while you are suggesting db-path19:04
jaypipesno it's a hack.19:04
enikanorov_db path was discussed and we have chosen another approach19:04
enikanorov_in fact we can't do lb_types (as you named it) a 'real resource'19:05
enikanorov_because such resource will depend on configuration, so we'll have to sync db state with configuration, which is not good19:06
*** aymenfrikha has joined #openstack-neutron19:06
jaypipesI think the main problem here is that the Neutron LBaaS API has no resource concept of a "load balancer", like the Atlas API does.19:06
enikanorov_how would it help? (btw: https://review.openstack.org/#/c/60207/ adds such entity)19:07
jaypipesbecause it would give a place to attach the "lb_type" attribute to.19:08
enikanorov_currently such place is 'pool'19:09
enikanorov_but it is orthogonal to a problem that we discuss19:09
jaypipessorry if I sounds annoyed... this patch (and strategy of doing this in a configuration file) is just so opposite to how pretty much *everything else* in OpenStack works.19:11
jaypipesand it just feels wrong.19:11
*** sweston has joined #openstack-neutron19:11
*** networkstatic has joined #openstack-neutron19:11
jaypipesenikanorov: was Mark McClain involved in the discussions about this at HK summit?19:12
enikanorov_that's funny, but it looks like neutron community had directly opposite opinion on the way of configuring providers19:12
jaypipesenikanorov: got a link I can read with the history there?19:12
enikanorov_jaypipes: it was introduced in Havana, and discussed at that time19:12
enikanorov_hmm, i would hard for me to find a link right now, it's more thatn 6 month old-thread19:13
enikanorov_the essence of the resolution is that we don't store providers and their configuration in db19:13
enikanorov_as a consequence, we not store any derived object (like flavor) in the db as well19:14
jaypipesthis isn't about storing driver configuration in the database!19:14
jaypipesthis is about having a way to indicate the type of load balancer that a user wants and how to communicate that information to the driver.19:14
enikanorov_currentl user is able to indicate a type of the balancer19:15
*** sweston has quit IRC19:15
jaypipesthen what precisely is the need to have a separate driver configuration section?19:15
enikanorov_user has all the tools to make a choice: list available providers, pass provider name to POST /pools19:16
jaypipesso what is the need for a separate "driver name" option in the configuration file?19:16
enikanorov_jaypipes: like said, fast-slow-cheap-broken services all managed by the same driver19:17
jaypipesyes? and why do we need driver_name configuration option and separate config sections for this?19:17
jaypipesit is provided as a parameter to create_pool()...19:17
jaypipesso why do we need separate configurion file sections.19:17
enikanorov_how else you provide 2 different sets of configuration to a driver?19:18
jaypipesit's all one set of configuration options.19:18
enikanorov_(in fact you'l have 2 instances of the same driver)19:18
jaypipesno... you have a single driver instance.19:19
*** clev has joined #openstack-neutron19:19
jaypipesthat does certain setups differently depending on the value of the provider parameter.19:19
jaypipesto create_pool...19:19
enikanorov_provider passed to create_pool directly maps to the driver currently19:19
jaypipesand that is the problem.19:20
enikanorov_we're not planning to extend rest api to introduce flavors there19:20
jaypipesit's mixing badly implementation with the API interface.19:20
enikanorov_API remains consistent, UX remains consistent19:20
enikanorov_I don't see a problem19:20
jaypipesAPi remains terrible. UX remains horrible.19:20
* jaypipes has never ever met anyone who likes Neutron API. sorry.19:21
enikanorov_well, i never met anyone who liked openstack,19:21
enikanorov_so what?19:21
jaypipeshmm, related?19:21
enikanorov_we have to work with what is at hand19:21
jaypipesno, we can make it better. instead of keeping bad interfaces.19:21
enikanorov_yes, we can, but first of all, we need to maintain backward compatibility of or sweet terrible API19:22
*** vkozhukalov has quit IRC19:22
enikanorov_that is the first point on the list19:22
enikanorov_*of our19:22
jaypipesI am starting to understand why the LB community is so fragmented in OpenStack... Atlas, Libra, Neutron...19:23
enikanorov_it's fragmented because all three were started simultaneously and we have never met together19:23
*** salv-orlando has joined #openstack-neutron19:24
enikanorov_jaypipes: introducing a single bit of API takes release cycle. getting rid of a bit of API takes 2 cycles19:25
jaypipesenikanorov: I am fully aware of this.19:25
enikanorov_implementing flavors with configuration is not only  simple, it goes inline with the initial design19:25
jaypipeswhat is this "initial design" of which you speak?19:25
enikanorov_the decision to go with configuration rather than db19:26
jaypipesand how many people other than you and Evgeny were involved in this discussion?19:26
jaypipesjust curious...19:26
enikanorov_cores, ptl, salv-orlando in particular (hi, salv-orlando!)19:26
salv-orlandowhat discussion? (if it's a long story I'll scroll the irc log)19:27
jaypipessalv-orlando: of using the configuration files of neutron.conf to implement "load balancer flavors".19:27
enikanorov_salv-orlando: we're fighting for flavors19:28
jaypipessalv-orlando: this hack: https://review.openstack.org/#/c/64139/19:28
enikanorov_i remember i personally has rewritten provider configuration patch 2 or 3 times going back and forth between db and configuration file19:29
salv-orlandoah *that* hack. I totally understand why it's been pushed, but I think we've accrued too much technical debt on this particular19:29
jaypipessigh... it's not configuring the driver, though...19:29
salv-orlandotopic, as enikanorov_ can confirm with his going back and forth19:30
enikanorov_configurting the driver is just a consequence19:30
*** markwash_ has joined #openstack-neutron19:30
enikanorov_we can't then ake db content be dependent on what is in configuration19:30
enikanorov_*make19:30
jaypipesthere's a driver. there's a single set of configuration options for that driver. the user calls POST /pools, the driver gets create_pool() called. part of that call's parameters is the "provider type"... just have the (single) driver do something different depending on the value of "provider"19:31
enikanorov_so lb_type, aka flavor can't be a resource19:31
*** harlowja is now known as harlowja_away19:31
enikanorov_jaypipes: yes, and we doing it, by mapping provider name to provider options in config19:31
jaypipesenikanorov: first, yes it can be a resource. but even if you don't want to implement it as a resource (and continue to use the "provider" as a hack for lb_type), you can still do all of this with a single driver and single set of configuration options.19:32
*** markwash has quit IRC19:32
*** markwash_ is now known as markwash19:32
jaypipesenikanorov: I have half a mind to just git review -d that patch and change it to show how the radware driver can just use one set of configurtation options.19:34
enikanorov_the bottom of the problem is that lb_type is db-resource, and lb_type depends on the provider, which is a configuration resource19:34
enikanorov_jaypipes: i think 10-15 lines of pseudocode will be well enough19:35
enikanorov_may be it could resolve misunderstanding between us :)19:35
jaypipesenikanorov: ok. I will do that then. code talks.19:35
jaypipesenikanorov: tomorrow, though... unfortunately, I have to complete documentation about setting up external testing platforms today ;)19:35
enikanorov_ok, sure19:36
*** sweston has joined #openstack-neutron19:41
*** markmcclain has joined #openstack-neutron19:46
*** amrit has joined #openstack-neutron19:50
*** notel has quit IRC19:50
*** markmcclain has quit IRC19:53
*** igorpkdmir has joined #openstack-neutron19:54
*** igorpkdmir has quit IRC19:54
*** markmcclain has joined #openstack-neutron19:54
*** izobrazilia has joined #openstack-neutron19:54
openstackgerritPaul Ward proposed a change to openstack/neutron: Disallow single IP subnets in neutron  https://review.openstack.org/6874219:54
*** markmcclain has quit IRC19:55
*** markmcclain has joined #openstack-neutron19:55
*** izobrazilia has quit IRC19:55
enikanorov_dkehn: hi19:56
dkehnenikanorov_: hi19:56
enikanorov_dkehn: can you explain -1 for https://review.openstack.org/#/c/55032/ ?19:56
enikanorov_ok, let me be more direct19:57
enikanorov_the jenkins failure in py26-27 jobs is not related to my patch19:57
dkehnjust doing a straight setup and then testing with tox -epy27 I see a failure, but when doing a isolated test it works19:57
enikanorov_it is going to be fixed by another commit19:57
*** zzelle has joined #openstack-neutron19:58
enikanorov_that one: https://review.openstack.org/#/c/67537/\19:58
dkehnenikanorov_: I see , didn't see the dependancy19:58
dkehnenikanorov_: should there be one there?19:58
enikanorov_half of neutron patches are failing py27 and py26 jobs19:58
*** harlowja_away is now known as harlowja19:59
enikanorov_because of the issue resolved by 6753719:59
dkehnenikanorov_: I'll take of the -1, k19:59
enikanorov_so there's no need to make a  dependency19:59
dkehnenikanorov_: better19:59
*** terrylhowe has joined #openstack-neutron20:00
*** ChuckC has joined #openstack-neutron20:00
*** ivar-lazzaro has quit IRC20:01
*** ChuckC has quit IRC20:01
*** sn6i23a has joined #openstack-neutron20:06
*** tongli has quit IRC20:06
*** dave_tucker is now known as dave_tucker_zzz20:13
*** rohit404 has quit IRC20:14
*** pcarver has quit IRC20:15
*** sweston_ has joined #openstack-neutron20:17
*** alex_klimov has quit IRC20:17
*** sweston has quit IRC20:18
*** rohit404 has joined #openstack-neutron20:18
*** DynamiteXXL has joined #openstack-neutron20:19
*** sweston_ has quit IRC20:22
*** aveiga has quit IRC20:29
*** markmcclain has quit IRC20:29
*** arash has quit IRC20:31
*** Guest67331 has quit IRC20:31
*** jdev789_away has joined #openstack-neutron20:32
*** jgrimm has quit IRC20:32
*** jdev789_away is now known as jdev78920:34
*** jdev789 has quit IRC20:35
*** sweston has joined #openstack-neutron20:37
*** sputnik13net has quit IRC20:40
openstackgerritCarl Baldwin proposed a change to openstack/neutron: L3 agent fetches the external network id once  https://review.openstack.org/6692820:41
*** sweston has quit IRC20:42
*** sweston has joined #openstack-neutron20:47
*** SumitNaiksatam has quit IRC20:51
*** marun has quit IRC20:51
*** SumitNaiksatam has joined #openstack-neutron20:51
*** marun has joined #openstack-neutron20:52
*** sweston has quit IRC20:52
*** markmcclain has joined #openstack-neutron20:54
*** sputnik13net has joined #openstack-neutron20:54
*** sweston has joined #openstack-neutron20:55
*** jgrimm has joined #openstack-neutron20:59
*** jecarey has quit IRC21:00
*** sweston has quit IRC21:07
*** sweston has joined #openstack-neutron21:09
*** jgrimm has quit IRC21:09
*** sputnik13net has quit IRC21:12
*** arash has joined #openstack-neutron21:15
*** dave_tucker_zzz is now known as dave_tucker21:15
*** markmcclain has quit IRC21:17
*** djoreilly has quit IRC21:19
*** salv-orlando has quit IRC21:20
*** marun has quit IRC21:24
*** alexpilotti has quit IRC21:25
*** alexpilotti has joined #openstack-neutron21:26
*** julim has quit IRC21:26
*** marun has joined #openstack-neutron21:27
*** mlavalle has quit IRC21:30
*** alexpilotti has quit IRC21:30
*** markmcclain has joined #openstack-neutron21:33
*** rohit404 has quit IRC21:37
*** terrylhowe has quit IRC21:49
*** mlavalle has joined #openstack-neutron21:58
*** sweston has quit IRC21:58
*** sweston has joined #openstack-neutron22:04
*** mfink has quit IRC22:04
*** thomasem has quit IRC22:05
*** sweston_ has joined #openstack-neutron22:07
*** sweston has quit IRC22:08
*** WackoRob_ has joined #openstack-neutron22:09
*** jecarey has joined #openstack-neutron22:10
*** WackoRobie has quit IRC22:12
*** WackoRob_ has quit IRC22:13
*** sweston_ has quit IRC22:14
*** sweston has joined #openstack-neutron22:17
*** marun has quit IRC22:20
*** pcm_ has quit IRC22:21
*** nati_ueno has quit IRC22:21
*** morganfainberg is now known as morganfainberg|z22:21
*** salv-orlando has joined #openstack-neutron22:22
*** networkstatic has quit IRC22:30
*** mfink has joined #openstack-neutron22:35
*** sn6i23a has quit IRC22:36
*** peristeri has quit IRC22:37
*** rms_13 has joined #openstack-neutron22:39
*** arash has quit IRC22:42
*** jroovers has quit IRC22:42
*** sweston_ has joined #openstack-neutron22:44
*** WackoRobie has joined #openstack-neutron22:45
*** sweston has quit IRC22:47
*** banix has quit IRC22:47
*** mlavalle has quit IRC22:48
openstackgerritIvar Lazzaro proposed a change to openstack/neutron: Embrane Tempest Compliance  https://review.openstack.org/6667122:54
*** gdubreui has joined #openstack-neutron22:56
*** sweston_ has quit IRC22:57
*** markmcclain has quit IRC22:57
*** gdubreui has quit IRC22:58
*** gdubreui has joined #openstack-neutron22:58
*** salv-orlando_ has joined #openstack-neutron23:00
*** salv-orlando has quit IRC23:03
*** salv-orlando_ is now known as salv-orlando23:03
*** thuc has quit IRC23:04
*** thuc has joined #openstack-neutron23:04
*** changbl has quit IRC23:05
*** thuc has quit IRC23:09
*** suresh12 has joined #openstack-neutron23:10
*** bjornar has joined #openstack-neutron23:10
openstackgerritSalvatore Orlando proposed a change to openstack/neutron: Introduce status for floating IPs  https://review.openstack.org/6686623:11
*** sweston has joined #openstack-neutron23:14
openstackgerritSalvatore Orlando proposed a change to openstack/neutron: Remove garbage in vim header  https://review.openstack.org/6879823:14
*** dims has quit IRC23:15
*** sweston has quit IRC23:18
*** zzelle has quit IRC23:24
*** dims has joined #openstack-neutron23:30
*** dave_tucker is now known as dave_tucker_zzz23:33
*** dave_tucker_zzz is now known as dave_tucker23:34
*** WackoRobie has quit IRC23:38
*** clev has quit IRC23:38
*** WackoRobie has joined #openstack-neutron23:39
*** MM_at_HP has joined #openstack-neutron23:43
MM_at_HPFeels like Friday23:43
*** morganfainberg|z is now known as morganfainberg23:48
*** markmcclain has joined #openstack-neutron23:52

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!