Monday, 2016-09-05

*** tflynn has quit IRC00:06
*** tflynn has joined #openstack-neutron00:07
*** sdake has joined #openstack-neutron00:07
*** tflynn has quit IRC00:19
*** tflynn has joined #openstack-neutron00:20
*** lujinluo has joined #openstack-neutron00:20
*** jckasper has joined #openstack-neutron00:22
*** amitkqed has joined #openstack-neutron00:26
*** jckasper has quit IRC00:27
*** mtaylor22 has joined #openstack-neutron00:34
*** tflynn has quit IRC00:34
*** hoangcx has joined #openstack-neutron00:35
*** clenimar has joined #openstack-neutron00:39
*** tflynn has joined #openstack-neutron00:39
*** sdake has quit IRC00:39
*** csmart has left #openstack-neutron00:41
*** chlong has joined #openstack-neutron00:45
*** clenimar has quit IRC00:45
*** mfuruta has joined #openstack-neutron00:51
*** markvoelker has joined #openstack-neutron00:52
openstackgerritKengo Hobo proposed openstack/neutron: Disallow specifying too long description for meter-label  https://review.openstack.org/35149300:54
*** markvoelker has quit IRC00:57
openstackgerritKengo Hobo proposed openstack/neutron: Handle label_id's DBReferenceError when creating label-rule  https://review.openstack.org/35959100:59
*** adriant has quit IRC01:03
*** puck has quit IRC01:04
*** SKipp_ has joined #openstack-neutron01:06
*** bigjools has quit IRC01:12
*** Guest99973 has joined #openstack-neutron01:12
*** rm_work has quit IRC01:12
*** rm_work has joined #openstack-neutron01:13
*** amotoki has joined #openstack-neutron01:13
*** clsqwerty has joined #openstack-neutron01:13
*** StevenK has quit IRC01:14
*** Guest62633 has quit IRC01:14
*** bigjools has joined #openstack-neutron01:16
*** gongysh has joined #openstack-neutron01:17
*** rossella_s has quit IRC01:18
*** dansmith has joined #openstack-neutron01:19
*** rossella_s has joined #openstack-neutron01:19
*** puck has joined #openstack-neutron01:19
*** salv-orlando has joined #openstack-neutron01:21
*** StevenK has joined #openstack-neutron01:22
*** salv-or__ has quit IRC01:24
*** tflynn has quit IRC01:24
*** fnaval has quit IRC01:25
openstackgerritKengo Hobo proposed openstack/neutron-vpnaas: Add router_id to query when check subnet is used by vpnservice  https://review.openstack.org/34883501:25
*** clenimar has joined #openstack-neutron01:31
*** stanzgy has joined #openstack-neutron01:35
*** wu_ng has joined #openstack-neutron01:38
*** EinstCrazy has joined #openstack-neutron01:43
*** EinstCrazy has quit IRC01:43
*** EinstCrazy has joined #openstack-neutron01:44
*** yulong|away has quit IRC01:48
*** tflynn has joined #openstack-neutron01:50
*** tflynn has quit IRC01:53
*** EinstCrazy has quit IRC01:53
*** EinstCrazy has joined #openstack-neutron01:54
*** StevenK has quit IRC01:55
*** StevenK has joined #openstack-neutron01:56
*** lilintan has joined #openstack-neutron02:00
*** thorst has joined #openstack-neutron02:02
*** mtaylor22 has quit IRC02:03
*** EinstCra_ has joined #openstack-neutron02:04
*** yedongcan has joined #openstack-neutron02:06
*** thorst has quit IRC02:07
*** EinstCrazy has quit IRC02:09
*** shihanzhang has quit IRC02:12
*** shihanzhang has joined #openstack-neutron02:13
*** banix has joined #openstack-neutron02:18
*** sticker has joined #openstack-neutron02:18
*** banix has quit IRC02:19
*** clsqwerty has quit IRC02:22
*** murphy_zhao has joined #openstack-neutron02:28
openstackgerritIWAMOTO Toshihiro proposed openstack/neutron: fullstack: Add security group tests  https://review.openstack.org/34132802:33
*** ociuhandu has quit IRC02:34
*** michchap has joined #openstack-neutron02:35
*** chlong has quit IRC02:36
openstackgerritIWAMOTO Toshihiro proposed openstack/neutron: Skip setup_port_filters for skipped_devices  https://review.openstack.org/35219602:40
*** liuyulong has joined #openstack-neutron02:40
*** fnaval has joined #openstack-neutron02:40
*** mtaylor22 has joined #openstack-neutron02:43
*** lilintan has quit IRC02:45
*** amotoki has quit IRC02:46
*** ociuhandu has joined #openstack-neutron02:53
*** markvoelker has joined #openstack-neutron02:53
*** chlong has joined #openstack-neutron02:53
*** yb has joined #openstack-neutron02:54
*** markvoelker has quit IRC02:57
*** thorst has joined #openstack-neutron03:02
*** thorst_ has joined #openstack-neutron03:03
*** amotoki has joined #openstack-neutron03:05
*** ociuhandu has quit IRC03:06
*** thorst has quit IRC03:06
*** thorst_ has quit IRC03:07
*** lilintan has joined #openstack-neutron03:10
*** amotoki has quit IRC03:13
openstackgerritZhengwei Gao proposed openstack/neutron: In case of: add_metering_label_rule fail to update iptables rule  https://review.openstack.org/36107203:25
*** annp has joined #openstack-neutron03:26
*** amitkqed has quit IRC03:27
*** amitkqed has joined #openstack-neutron03:28
*** lilintan has quit IRC03:32
*** stanzgy has quit IRC03:35
*** lilintan has joined #openstack-neutron03:44
*** fnaval has quit IRC03:44
*** amotoki has joined #openstack-neutron04:01
*** Alex_Stef has joined #openstack-neutron04:03
*** clenimar has quit IRC04:03
*** clenimar has joined #openstack-neutron04:05
*** clenimar has quit IRC04:13
*** jckasper has joined #openstack-neutron04:24
openstackgerritlilintan proposed openstack/neutron: Use method is_valid_ipv* from oslo.utils  https://review.openstack.org/36543304:25
*** moshele has joined #openstack-neutron04:25
*** jckasper has quit IRC04:29
*** ilyashakhat_mobi has joined #openstack-neutron04:30
*** oshvartz has quit IRC04:30
*** claudiub has joined #openstack-neutron04:33
*** ilyashakhat_mobi has quit IRC04:36
*** obondarev has joined #openstack-neutron04:39
*** trananhkma has joined #openstack-neutron04:47
*** wolverineav has joined #openstack-neutron04:49
*** links has joined #openstack-neutron04:49
*** oanson has joined #openstack-neutron04:53
*** markvoelker has joined #openstack-neutron04:53
*** tflynn has joined #openstack-neutron04:55
*** markvoelker has quit IRC04:58
*** garyk has joined #openstack-neutron04:58
*** tflynn has quit IRC05:01
*** buttercup has joined #openstack-neutron05:02
*** ekuris has joined #openstack-neutron05:04
*** bkopilov has joined #openstack-neutron05:06
*** bkopilov has quit IRC05:06
*** bkopilov has joined #openstack-neutron05:07
*** fragatina has quit IRC05:07
garykkevinbenton: around?05:14
openstackgerritZhengwei Gao proposed openstack/neutron: In case of: add_metering_label_rule fail to update iptables rule  https://review.openstack.org/36107205:16
*** murphy_zhao has quit IRC05:19
*** murphy_zhao has joined #openstack-neutron05:23
*** dkehn has quit IRC05:27
*** dkehn_ has quit IRC05:28
*** aqkhan__ has joined #openstack-neutron05:33
*** clenimar has joined #openstack-neutron05:33
*** dkehn has joined #openstack-neutron05:34
*** aqkhan_ has quit IRC05:34
*** clsqwerty has joined #openstack-neutron05:35
*** aqkhan has joined #openstack-neutron05:35
*** kobis has joined #openstack-neutron05:37
*** aqkhan__ has quit IRC05:38
*** padkrish has joined #openstack-neutron05:38
*** dkehn_ has joined #openstack-neutron05:40
*** padkrish has quit IRC05:40
*** puck has quit IRC05:41
*** jhershbe has joined #openstack-neutron05:47
*** jhershbe_ has joined #openstack-neutron05:47
*** matrohon has quit IRC05:48
*** puck has joined #openstack-neutron05:52
*** djan_ has joined #openstack-neutron05:53
*** djan has quit IRC05:55
*** vikram has joined #openstack-neutron05:55
*** oshvartz has joined #openstack-neutron05:58
*** iranzo has joined #openstack-neutron06:01
*** amotoki has quit IRC06:05
*** mfuruta has quit IRC06:07
*** mfuruta has joined #openstack-neutron06:07
*** vikram has quit IRC06:07
openstackgerritZhengwei Gao proposed openstack/neutron: Preventing iptables rule to be thrashed  https://review.openstack.org/36401906:11
*** clsqwerty has quit IRC06:17
*** prateek has joined #openstack-neutron06:18
openstackgerritQunyingRan proposed openstack/neutron: Remove used ip which not in subnet pool from network used ip statistics  https://review.openstack.org/32238706:19
*** Alex_Stef has quit IRC06:20
*** pcaruana has joined #openstack-neutron06:20
*** vikram has joined #openstack-neutron06:22
*** wolverineav has quit IRC06:24
*** buttercup has quit IRC06:27
*** thorst has joined #openstack-neutron06:33
*** wolverineav has joined #openstack-neutron06:38
*** slaweq_ has joined #openstack-neutron06:39
openstackgerritgaryk proposed openstack/neutron: Trunk: remove database transactions from the plugin  https://review.openstack.org/36545906:39
jschwarzliuyulong, hey06:41
jschwarzliuyulong, re: https://review.openstack.org/#/c/265672/06:41
*** wolverineav has quit IRC06:42
*** roeyc has joined #openstack-neutron06:45
*** nyechiel has joined #openstack-neutron06:45
*** amotoki has joined #openstack-neutron06:46
openstackgerritgaryk proposed openstack/neutron: Fix InvalidInput exception output  https://review.openstack.org/36364106:47
*** amotoki has quit IRC06:48
*** amotoki has joined #openstack-neutron06:48
*** slaweq has quit IRC06:49
*** lihi has joined #openstack-neutron06:52
*** chlong has quit IRC06:53
*** markvoelker has joined #openstack-neutron06:54
*** slaweq has joined #openstack-neutron06:56
*** ijw has joined #openstack-neutron06:57
*** tflynn has joined #openstack-neutron06:57
*** ijw has quit IRC06:57
*** ijw has joined #openstack-neutron06:58
*** markvoelker has quit IRC06:59
*** andreas_s has joined #openstack-neutron06:59
*** roeyc has quit IRC06:59
*** edand has joined #openstack-neutron07:00
*** nmagnezi_ has joined #openstack-neutron07:00
*** stanzgy has joined #openstack-neutron07:01
*** tflynn has quit IRC07:02
*** kobis has quit IRC07:02
*** tmorin has joined #openstack-neutron07:03
*** obondarev has quit IRC07:03
*** Guest99973 has quit IRC07:03
liuyulongjschwarz, hi07:04
*** xiaoyang has joined #openstack-neutron07:04
jschwarzliuyulong, see my comment on the patch :)07:04
liuyulongjschwarz, OK07:04
openstackgerritMerged openstack/neutron: Set trunk status to DOWN when parent port is unbound  https://review.openstack.org/36148107:05
*** wolverineav has joined #openstack-neutron07:05
*** chlong has joined #openstack-neutron07:06
liuyulongjschwarz, Oh, I see, without that patch the new trace I got that trace, that patch is to fix it.07:06
*** sleviim has joined #openstack-neutron07:06
jschwarzliuyulong, I see07:07
jschwarzliuyulong, so please open a new bug with server traces, etc07:07
liuyulongjschwarz, not a new bug.07:07
*** abregman has joined #openstack-neutron07:08
liuyulongjschwarz, It's the same bug. The trace was already pasted in the LP bug 1607381 description.07:09
openstackLaunchpad bug 1607381 in neutron "HA router in l3 dvr_snat/legacy agent has no ha_port" [Undecided,In progress] https://launchpad.net/bugs/1607381 - Assigned to LIU Yulong (dragon889)07:09
*** liuyulong has left #openstack-neutron07:09
*** liuyulong has joined #openstack-neutron07:09
*** wolverineav has quit IRC07:10
liuyulongjschwarz, `infinite loop` caused by HA router without ha_port.07:10
openstackgerritgaryk proposed openstack/neutron: ML2: improve performance of get_networks  https://review.openstack.org/30226607:10
*** ilyashakhat_mobi has joined #openstack-neutron07:11
liuyulongjschwarz, It's really an old story: "L3 agent still has chance to get a HA router without ha_port".07:12
*** mtaylor22 has quit IRC07:13
*** wolverineav has joined #openstack-neutron07:14
*** jpena|off is now known as jpena07:17
*** yfried has joined #openstack-neutron07:18
*** wolverineav has quit IRC07:19
*** ilyashakhat_mobi has quit IRC07:19
openstackgerritgaryk proposed openstack/neutron: L3: ensure that self.l3_rpc_notifier set and not None  https://review.openstack.org/36547207:20
*** salv-orl_ has joined #openstack-neutron07:20
*** devvesa has joined #openstack-neutron07:23
*** salv-orlando has quit IRC07:23
*** Alex_Stef has joined #openstack-neutron07:25
*** vikram has quit IRC07:27
jschwarzliuyulong, is this on master?07:29
jschwarzliuyulong, it is important we understand what happens server-side that triggers this behaviour.07:29
*** abregman_ has joined #openstack-neutron07:30
*** sleviim has left #openstack-neutron07:31
*** sleviim has joined #openstack-neutron07:31
*** sleviim1 has joined #openstack-neutron07:32
*** abregman has quit IRC07:33
*** tesseract- has joined #openstack-neutron07:33
*** oshvartz has quit IRC07:34
*** rubasov has joined #openstack-neutron07:34
liuyulongjschwarz, It's stable/mitaka, with fix of bug 1606844.07:35
openstackbug 1606844 in neutron "L3 agent constantly resyncing deleted router" [High,Fix released] https://launchpad.net/bugs/1606844 - Assigned to Oleg Bondarev (obondarev)07:35
*** trananhkma has quit IRC07:35
*** wolverineav has joined #openstack-neutron07:35
jschwarzliuyulong, ah, a very important detail you didn't mention anywhere :)07:35
jschwarzliuyulong, can you check this on the latest master?07:35
*** namnh has joined #openstack-neutron07:36
*** obondarev has joined #openstack-neutron07:36
*** hynekm has joined #openstack-neutron07:37
liuyulongjschwarz, We only have a 40 nodes mitaka envrionment. and that trace was got after run the following rally task: http://paste.openstack.org/show/566802/. And I don't think a local VM deployment could handle it.07:39
*** ijw has quit IRC07:40
*** ijw has joined #openstack-neutron07:40
jschwarzliuyulong, ok07:41
jschwarzliuyulong, so lets narrow it down - please dig into the server code and try to figure out exactly which objects are conflicting07:41
jschwarzliuyulong, we can probably just backport one of the fixes to mitaka07:41
*** vikram has joined #openstack-neutron07:43
*** wolverineav has quit IRC07:44
openstackgerritThomas Bechtold proposed openstack/neutron: Don't include [agent] section in neutron.conf.sample  https://review.openstack.org/36548807:47
liuyulongjschwarz, for now, could neutron master branch code ensure that the L3 agent never has chance to get HA router without ha_port ?07:49
*** vikram has quit IRC07:49
jschwarzliuyulong, nothing could "ensure" that 100%, but master is better maintained IMO07:50
liuyulongjschwarz, I noticed that you abandoned some changes.07:50
jschwarzliuyulong, I did, because we merged https://review.openstack.org/#/c/364278/ so the rest aren't needed07:50
openstackgerritHynek Mlnarik proposed openstack/neutron: API tests: Check MTU sanity of trunk/subport  https://review.openstack.org/36420807:51
*** bjolo has quit IRC07:51
*** matrohon has joined #openstack-neutron07:52
liuyulongjschwarz, OK, i will test the master with some lower rally pressure.07:52
jschwarzliuyulong, thanks, but what we really need is to figure out what's going on in the server-side on stable/mitaka.07:53
*** oshvartz has joined #openstack-neutron07:53
*** rdo_ has joined #openstack-neutron07:53
liuyulongjschwarz, I think it's still the race condition between HA router deleting and sync_ha_router_info.07:54
*** chenyan has joined #openstack-neutron07:54
jschwarzliuyulong, that doesn't make sense since the router is ALLOCATING so it shouldn't send it07:54
jschwarzliuyulong, can you please make sure?07:54
*** fzdarsky has joined #openstack-neutron07:56
*** rmart04 has joined #openstack-neutron07:56
*** Alex_Stef has quit IRC07:57
liuyulongjschwarz, https://review.openstack.org/#/c/265672/ patch set 10 I leave to Kevin, did that code calling tree have any change in master or mitaka ?07:57
*** nplanel has joined #openstack-neutron07:57
*** mickeys has joined #openstack-neutron07:58
*** tflynn has joined #openstack-neutron07:58
*** thorst has quit IRC07:58
*** mickeys has quit IRC07:59
*** zzzeek has quit IRC08:00
liuyulong`a fully-built (active) HA router is free to delete`08:00
jschwarzliuyulong, it doesn't work like that in neither mitaka or master08:00
*** zzzeek has joined #openstack-neutron08:01
jschwarzliuyulong, first the HA attributes are created, then they are bound to agents during the scheduling, and only then the router becomes active (after allocating)08:01
jschwarzliuyulong, so..08:01
jschwarzliuyulong, and when the router is allocating it isn't returned to the agents during get_ha_router_info (it's filtered out)08:01
*** jlibosva has joined #openstack-neutron08:02
*** tflynn has quit IRC08:02
openstackgerritIWAMOTO Toshihiro proposed openstack/neutron: Skip setup_port_filters for skipped_devices  https://review.openstack.org/35219608:03
liuyulongRouter is not shecduled to L3 agent when router is allocating state. And if it is scheduled, then L3 agent will sync router info. At this time, router was concurrently deleting due to a delete API.08:05
liuyulongjschwarz, ^08:06
jschwarzliuyulong, https://github.com/openstack/neutron/blob/stable/mitaka/neutron/db/l3_hamode_db.py#L44808:07
*** aloga has quit IRC08:07
jschwarzliuyulong, this clearly creates the HA router as ALLOCATING, schedules it on line 464 and then sets it to active on line 465 (and then sends a notification to the agents in line 468)08:07
*** saggi has joined #openstack-neutron08:09
*** prateek has quit IRC08:10
jschwarzliuyulong, also, https://github.com/openstack/neutron/blob/stable/mitaka/neutron/db/l3_agentschedulers_db.py#L368 makes sure that the get_sync doesn't return any router that is currently ALLOCATING08:10
*** Alex_Stef has joined #openstack-neutron08:11
liuyulongjschwarz, yes, it is. lets change to another perspective, when L3 agent sync a router info, is it ACTIVE state?08:12
*** ababich has quit IRC08:12
jschwarzliuyulong, it might but it might not08:12
*** ilyashakhat_mobi has joined #openstack-neutron08:13
liuyulongjschwarz, https://github.com/openstack/neutron/blob/stable/mitaka/neutron/db/l3_hamode_db.py#L448 this line just create DB stuffs, not notify L3 agent a router is coming to you. Line 468 will do that nitifacation. right?08:15
jschwarzliuyulong, right08:16
liuyulongjschwarz, It's now ACTIVE state, and then L3 agent sync the router info. RPC will go to https://github.com/openstack/neutron/blob/stable/mitaka/neutron/db/l3_hamode_db.py#L682.08:17
*** mosulica has joined #openstack-neutron08:17
liuyulongjschwarz, deleting API will goto https://github.com/openstack/neutron/blob/stable/mitaka/neutron/db/l3_hamode_db.py#L57408:18
jschwarzliuyulong, uhm yes08:18
liuyulongjschwarz, Things now get interesting then.08:18
jschwarzliuyulong, now I understand the scenario you had in mind08:18
jschwarzliuyulong, this was a good explaination, you should have begun with that :)08:18
*** openstackgerrit has quit IRC08:18
*** openstackgerrit has joined #openstack-neutron08:19
*** bfernando has joined #openstack-neutron08:21
*** xiaoyang has quit IRC08:21
*** xiaoyang has joined #openstack-neutron08:22
*** ushkalim has joined #openstack-neutron08:24
liuyulongjschwarz, https://github.com/openstack/neutron/blob/stable/mitaka/neutron/db/l3_hamode_db.py#L723, here sync RPC only get the basic info of the router, and then sync_ha_data will get nothing, because the router is concurrently deleted, as well the HA ports.08:25
jschwarzliuyulong, sounds reasonable. Please post this information in your patch :)08:26
*** jckasper has joined #openstack-neutron08:26
liuyulongjschwarz, the commit message has some, seems not enough, : )08:27
jschwarzliuyulong, also the last comments you made are confusing ;-)08:27
*** lucas-dinner is now known as lucasagomes08:28
*** salv-orlando has joined #openstack-neutron08:29
liuyulongjschwarz, its also one scenario of race conditon - "creating and deleting".08:30
*** chlong has quit IRC08:31
jschwarzliuyulong, yes, but a lot happens during the creation and deletion08:31
*** wolverineav has joined #openstack-neutron08:31
*** jckasper has quit IRC08:31
jschwarzliuyulong, it's one step away from me telling you "there's a race in Neutron when the server starts and I issue a request to create a resource"08:31
jschwarzliuyulong, very hard to figure out from this exactly what went wrong, don't you agree?08:31
liuyulongjschwarz, OK, you got me.08:32
*** salv-orl_ has quit IRC08:32
liuyulongjschwarz, : )08:32
liuyulongjschwarz, I will update the commit message.08:32
*** abregman_ has quit IRC08:33
liuyulongjschwarz, so, master branch seems also faces the same thing.08:33
jschwarzliuyulong, sounds reasonable08:34
*** thorst has joined #openstack-neutron08:34
jschwarzliuyulong, the only bug though is that the agent can't handle it properly :)08:34
*** devvesa has quit IRC08:35
liuyulongjschwarz, yes, that patch is done for that.08:35
*** wolverineav has quit IRC08:35
*** devvesa has joined #openstack-neutron08:36
*** liuyulong is now known as liuyulong|away08:38
*** bjolo has joined #openstack-neutron08:40
*** Guest30374 is now known as bradjones08:41
*** bradjones has joined #openstack-neutron08:41
*** thorst has quit IRC08:41
*** ushkalim has quit IRC08:42
*** abregman_ has joined #openstack-neutron08:46
*** obondarev has quit IRC08:47
*** wolverineav has joined #openstack-neutron08:47
*** trinaths has joined #openstack-neutron08:47
slaweq_amotoki: hello08:47
slaweq_I have question about Your comments in https://review.openstack.org/#/c/358344/4/api-ref/source/v2/quotas.inc08:47
amotokislaweq_: hello08:48
slaweq_I'm changing it according to Your comments08:48
slaweq_but I'm not sure if I should change order of error response codes in all methods or only in this one for default quotas?08:48
*** clsqwerty has joined #openstack-neutron08:48
amotokislaweq_: sorry. i misread your sentence. You just added 'default quota' first.08:49
slaweq_and same with network object in parameters.yaml - should I add only networks or all missing parameters?08:49
slaweq_amotoki: I added description of default quotas08:49
slaweq_and You wrote me in L61 in quotas.inc file: "It is better to be sorted by response code." - but it is in same orded in all methods described in this file08:50
slaweq_so should I change it only for my new description of get default quota method or change also for other?08:51
*** clsqwerty has quit IRC08:51
slaweq_or maybe I should fix other in another commit?08:51
slaweq_what would be best solution for You?08:51
*** wolverineav has quit IRC08:51
amotokislaweq_: re 'reponse code order' , there is no strict guideline for additive change. You can follow the guideline for the new contents or can fix all response code.08:52
*** djan_ has quit IRC08:52
amotokislaweq_: i am fine with either.08:52
*** djan has joined #openstack-neutron08:52
slaweq_ok, and what about quota-network parameter? should I change it in all methods or only in new one?08:53
amotokislaweq_: regarding 'network object', I see several attributes has wrong description in the URL in the comment L.76.08:54
*** wolverineav has joined #openstack-neutron08:54
amotokislaweq_: could you make similar change for all related objects?08:54
slaweq_yes, I can :)08:54
amotokislaweq_: thanks!08:54
slaweq_so I will do it in this patch for all, thx for help08:54
*** markvoelker has joined #openstack-neutron08:55
*** wolverineav has quit IRC08:58
*** markvoelker has quit IRC09:00
*** yfried has quit IRC09:00
slaweq_amotoki: one more question09:02
*** nplanel has quit IRC09:02
slaweq_as I see in https://github.com/openstack/neutron-lib/blob/master/api-ref/source/v2/parameters.yaml#L244309:02
slaweq_there is parameter "network_1" instead of "quota-network" which You suggested09:03
*** yfried has joined #openstack-neutron09:03
slaweq_should I use it?09:03
slaweq_I think it should be fine09:03
amotokislaweq_: yeah. a lot of <attribute>_N were generated automatically by the rst migration script.09:03
amotokislaweq_: I would suggest to rename 'network_1' to 'quota-network'.09:04
slaweq_so should I use it like it is or rename?09:04
*** openstackgerrit has quit IRC09:04
slaweq_ok09:04
slaweq_thx09:04
amotokislaweq_: i believe it imporves readability a lot :)09:04
*** clsqwerty has joined #openstack-neutron09:05
*** openstackgerrit has joined #openstack-neutron09:05
slaweq_yes, I agree09:05
*** ivc_ has joined #openstack-neutron09:05
*** ilyashakhat_mobi has quit IRC09:10
*** eezhova has joined #openstack-neutron09:16
*** djan has quit IRC09:17
*** sambetts|afk is now known as sambetts09:18
*** rossella_s has quit IRC09:18
*** rossella_s has joined #openstack-neutron09:19
*** wu_ng has quit IRC09:19
*** john-davidge has joined #openstack-neutron09:21
openstackgerritIhar Hrachyshka proposed openstack/neutron-vpnaas: Add support for Guru Meditation Reports for VPNaaS agents  https://review.openstack.org/28780109:21
*** donghao has joined #openstack-neutron09:22
jschwarzakamyshnikova, hey09:22
akamyshnikovajschwarz, hi09:22
jschwarzakamyshnikova, got a minute to talk about our lord and saviour DB errors? :)09:22
akamyshnikovajschwarz, :D09:22
akamyshnikovajschwarz, always :)09:23
jschwarzakamyshnikova, http://pastebin.com/jAKtZScb09:23
jschwarzakamyshnikova, any ideas? :)09:23
jschwarzakamyshnikova, this is liberty btwe09:24
jschwarzbtw*09:24
*** wolverineav has joined #openstack-neutron09:24
akamyshnikovajschwarz, deadlock :O09:24
jschwarzakamyshnikova, also, http://pastebin.com/YvhVMWQw09:24
jschwarzakamyshnikova, we see this in 2 different flows (but always triggered from add_ha_port)09:25
jschwarzakamyshnikova, /me is not happy this morning :<09:25
*** john-davidge has quit IRC09:25
*** john-dav_ has joined #openstack-neutron09:25
*** thorst has joined #openstack-neutron09:26
*** lujinluo has quit IRC09:26
*** lujinluo has joined #openstack-neutron09:26
jschwarzlujinluo, hey09:27
jschwarzlujinluo, we ended up merging that patch from last week so https://review.openstack.org/#/c/289190/ can be abandonded09:27
jschwarzlujinluo, (https://review.openstack.org/364278)09:27
akamyshnikovajschwarz, this was just ordinary work or rally test run?09:27
jschwarzakamyshnikova, it was a poor-mans' create-and-immediately-delete in a loop09:28
jschwarzakamyshnikova, i.e. neutron router-create; neutron router-delete for ~5 times09:28
jschwarzakamyshnikova, 3 controllers in HA09:28
*** wolverineav has quit IRC09:28
akamyshnikovajschwarz, seems that we need a retry for add_ha_port09:29
*** mfuruta has quit IRC09:29
amotokigaryk: ping09:29
jschwarzakamyshnikova, this was on our downstream (not upstream stable/liberty)09:30
jschwarzakamyshnikova, maybe it doesn't reproduce on stable/liberty and it's just our fault09:30
jschwarzakamyshnikova, I've never seen that error though, "ResourceClosedError"09:30
jschwarzakamyshnikova, when does it pop?09:30
*** yb has quit IRC09:32
kevinbentonrossella_s: ping09:32
*** donghao has quit IRC09:32
akamyshnikovajschwarz, actually I also haven't seen it appeared, probably something wrong on backend side09:33
*** prateek has joined #openstack-neutron09:33
akamyshnikovajschwarz, or something goes  very wrong in concurrent09:34
jschwarzakamyshnikova, looking at logs09:34
*** clsqwerty has quit IRC09:34
jschwarzakamyshnikova, mysqld.log doesn't show anything out of the ordinary09:35
*** wolverineav has joined #openstack-neutron09:36
garykamotoki: hi, whats up?09:37
amotokigaryk: is there any reason you didn't +W to neutron-lib patches?09:37
garykamotoki: i think that only armando and henry can approve09:37
garyki think that this is a way of enforcing that we do not break neutreon-lib (not sure i agree with it)09:38
amotokigaryk: let me check.09:38
akamyshnikovajschwarz, then I guess the was another race condition, does the add_ha_port is very different from what we have in stable liberty?09:40
amotokigaryk: I remember only neutron-drivers can approve patches of neutron-lib.09:40
jschwarzakamyshnikova, afair, nope09:40
jschwarzakamyshnikova, let me look again09:40
amotokigaryk: sorry for interrupt09:40
*** wolverineav has quit IRC09:41
garykamotoki: ok. np.09:42
garykamotoki: i had my license revoked :)09:42
jschwarzakamyshnikova, so we use create_object_with_dependency to surround add_ha_port09:43
jschwarzakamyshnikova, (upstream, we only started doing that in mitaka, not liberty09:43
jschwarz)09:43
akamyshnikovajschwarz, this should not be bad09:44
jschwarzakamyshnikova, I agree09:44
*** davidsha has joined #openstack-neutron09:45
akamyshnikovajschwarz, and on mikata based/master it does not appear?09:46
jschwarzakamyshnikova, I don't think so09:47
jschwarzakamyshnikova, I didn't check in a while though :\09:47
*** amotoki has quit IRC09:47
akamyshnikovajschwarz, how well does it reproduced? everytime or once?09:53
jschwarzakamyshnikova, pretty heavily actually09:53
*** nplanel has joined #openstack-neutron09:54
jschwarzakamyshnikova, let me run it a few times and I'll get back to you09:54
akamyshnikovajschwarz, :( let me check this for mitaka09:54
*** slaweq_ has quit IRC09:55
*** slaweq_ has joined #openstack-neutron09:56
*** sdake has joined #openstack-neutron09:57
*** donghao has joined #openstack-neutron09:58
*** ijw has quit IRC09:58
*** grauzikas has joined #openstack-neutron09:59
jschwarzakamyshnikova, ack, much appreciated09:59
*** sdake has quit IRC10:01
*** kobis has joined #openstack-neutron10:01
jschwarzakamyshnikova, we're getting some StaleDataError pretty consistently on L3HARouterAgentPortBinding but I know why that's happening10:02
jschwarzakamyshnikova, the rest haven't reproduced yet10:02
jschwarzakamyshnikova, I spoke to soon - one of them just did :)10:02
rossella_skevinbenton, pong10:04
*** amotoki has joined #openstack-neutron10:05
kevinbentonrossella_s: i have an OVO issue10:05
akamyshnikovajschwarz, how lucky you are :)10:05
rossella_skevinbenton, what kind of issue?10:05
kevinbentonrossella_s: these fields https://github.com/openstack/neutron/blob/86f7cfefb0c253781a9a1d9b417a2043531837f2/neutron/objects/extensions/standardattributes.py#L18-L1910:06
kevinbentonrossella_s: contain microseconds10:06
kevinbentonrossella_s: but the database does not store microseconds10:06
*** sdake has joined #openstack-neutron10:07
kevinbentonrossella_s: so if you create an object, stash it in the db, load a second copy from the DB and compare them, they will be different because of the timestamp10:07
*** gongysh has quit IRC10:08
rossella_skevinbenton, I see...we need to fix something in the way we handle standard attributes in the ovo code, it's probably easy...10:08
rossella_sjlibosva, you wrote that part, right?10:08
kevinbentonrossella_s: well basically the issue is that we can't use the upstream datetime type10:08
kevinbentonrossella_s: DateTimeField10:09
kevinbentonrossella_s: it's to_primitive doesn't truncate10:09
rossella_skevinbenton, so you have a fix already? :p10:09
kevinbentonrossella_s: so i can push an alternate type10:10
kevinbentonrossella_s: that truncates on serialization10:11
kevinbentonrossella_s: but i just wanted to verify that would be an ok approach10:11
rossella_skevinbenton, just a question if the db ignore microseconds, why when loading the second copy the object is different?10:12
kevinbentonrossella_s: loading the second copy would have .0000 for the microseconds10:13
kevinbentonrossella_s: where the original one would have whatever came from the creating caller10:13
rossella_skevinbenton, then I think your approach makes sense10:13
kevinbentonrossella_s: ok. i'll push up a WIP for you to take a look at10:14
kevinbentonrossella_s: how much longer are you online?10:14
*** amotoki has quit IRC10:14
rossella_skevinbenton, it's my day time...you have 6 or more hours...but it's night for you10:14
kevinbentonrossella_s: it's a holiday here, i'm allowed to stay up late! :)10:15
rossella_skevinbenton, oh my!10:15
*** hoangcx has quit IRC10:17
openstackgerritHynek Mlnarik proposed openstack/neutron: Check MTU sanity of trunk port subports  https://review.openstack.org/36291710:17
jschwarzakamyshnikova, lol :)10:19
openstackgerritHynek Mlnarik proposed openstack/neutron: Check MTU sanity of trunk port subports  https://review.openstack.org/36291710:20
kevinbentonjschwarz: always expect staledata10:21
kevinbentonjschwarz: we should have retry decorators around things that can throw that10:21
jschwarzkevinbenton, practically everything can throw that though ;-(10:22
*** ushkalim has joined #openstack-neutron10:22
kevinbentonjschwarz: sqlalchemy has some unexpected behavior when it comes to how long things are cached in the session10:22
*** rmart04 has quit IRC10:22
kevinbentonjschwarz: so you can even be working with old data if you call get_whatever directly before using it10:22
kevinbentonjschwarz, akamyshnikova: interesting reading :) https://review.openstack.org/#/c/364301/10:23
openstackgerritIhar Hrachyshka proposed openstack/neutron-lib: Add support for upper-constraints.txt for all tox targets  https://review.openstack.org/36556010:23
openstackgerritInessa Vasilevskaya proposed openstack/neutron: ovsfw: fix troublesome port_rule_masking  https://review.openstack.org/35378210:24
*** rmart04 has joined #openstack-neutron10:25
ivasilevskayajlibosva, ping :)10:25
kevinbentonjschwarz: the resourceclosederror should have been covered by https://github.com/openstack/neutron/commit/3213dd884e15d1e2b2209aed3aed82cd0213a677 i believe10:26
*** EinstCra_ has quit IRC10:26
kevinbentonjschwarz: so it may need a back-port10:26
kevinbentonjschwarz: (again assuming this thing is protected by a decorator somewhere)10:26
jlibosvaivasilevskaya: heya :)10:29
*** nplanel has quit IRC10:29
openstackgerritSlawek Kaplonski proposed openstack/neutron-lib: Add info about API method to get default quotas  https://review.openstack.org/35834410:30
jschwarzkevinbenton, https://review.openstack.org/#/c/364301/ is interesting indeed10:30
*** ramishra has quit IRC10:30
*** rmart04_ has joined #openstack-neutron10:30
ivasilevskayajlibosva, I had a hard week writing documentation and iwamoto had a tough time reviewing it. If you take a look at https://review.openstack.org/353782 and say if it's ok, we would be grateful10:30
*** rmart04 has quit IRC10:31
*** rmart04_ is now known as rmart0410:31
ivasilevskayaThat's port_masking for ovsfw patch10:31
jschwarzkevinbenton, so basically everything is retriable since that commit?10:31
*** namnh has quit IRC10:32
kevinbentonjschwarz: since which commit?10:32
kevinbentonthe 3213dd88 one?10:32
*** ramishra has joined #openstack-neutron10:32
jschwarzkevinbenton, yes10:33
kevinbentonjschwarz: all that does is includes resourceclosederror in the list of exceptions the retry decorators sitting in the API layer think are okay10:33
kevinbentonjschwarz: to retry10:33
*** trinaths has left #openstack-neutron10:33
kevinbentonjschwarz: the issue is that if the exceptions aren't allowed to bubble back up the API layer, the retry decorator won't see them10:34
jschwarzkevinbenton, aye10:34
jschwarzkevinbenton, so I'm seeing some other weird stuff10:34
kevinbentonjschwarz: so as long as the resourceclosederror goes uncaught10:34
kevinbentonjschwarz: it would be retried after that commit10:34
jschwarzkevinbenton, specifically http://pastebin.com/xDVSyrfv10:35
jlibosvaivasilevskaya: I have it on my todo for today. I briefly looked in the morning that the documentation improved a lot :) Good job10:36
jschwarzkevinbenton, also there was some weird occurance of an error someplace in the loop, and after that error none of the create_router() calls succeeded10:36
jschwarzkevinbenton, I'm gonna apply that patch and turn on debug mode, see what happens then10:37
jlibosvaivasilevskaya: I only worry about correctness of algorithm and as there are special cases, whether we covered all those special cases. But maybe if I review with a bigger detail, I'll get my answers10:37
kevinbentonjschwarz: the old-school mysql-C connector deadlocks might not be caught by the decorator10:37
kevinbentonjschwarz: oh, nevermind. that should be ok. it's identified as a deadlock10:38
jschwarzkevinbenton, so basically it's a magic "catch all deadlocks" decorator? :)10:38
*** sdake has quit IRC10:38
kevinbentonjschwarz: it's a "catch all db transient issues/race conditions" decorator10:39
kevinbentonjschwarz: so staledata, duplicaterecord, etc10:39
jschwarzkevinbenton, that originate from the API (as opposed from the rpc)10:39
jschwarzright?10:39
kevinbentonjschwarz: RPC should have them too10:39
jschwarzkevinbenton, ah yes, here it is10:40
kevinbentonjschwarz: https://github.com/openstack/neutron/blob/master/neutron/api/rpc/handlers/l3_rpc.py#L79-L8010:40
jschwarzkevinbenton, plans to backport them to mitaka or liberty?10:40
lujinluojschwarz, yeah, i received the email. thank you for the efforts! i am on my way to abandon that patch set ;)10:40
kevinbentonjschwarz: https://github.com/openstack/neutron/blob/stable/liberty/neutron/api/rpc/handlers/l3_rpc.py#L64-L6510:41
ivasilevskayajlibosva, there is a paragraph about correctness proof and all corner cases I could think of are covered with tests. Will be waiting for your feedback then :)10:41
kevinbentonjschwarz: been around a long time :)10:41
*** ivasilevskaya has left #openstack-neutron10:41
*** ivasilevskaya has joined #openstack-neutron10:41
jschwarzkevinbenton, must have not been looking at the right place :)10:41
*** chenyan has quit IRC10:41
kevinbentonjschwarz: however, if you aren't using pymysql, some of the deadlock exceptions are different10:41
kevinbentonjschwarz: and the decorator doesn't catch them10:42
kevinbentonjschwarz: _mysql_exceptions.OperationalError implies that you are using the mysql-C connector10:42
*** ihrachys has joined #openstack-neutron10:43
jschwarzkevinbenton, aye10:43
ihrachysjlibosva: I am good with your arguments in https://review.openstack.org/#/c/364880/10:43
kevinbentonjschwarz: but i do see that it shows deadlock to the left, so it might be fine10:43
ihrachysjlibosva: wanna respin for the minor clarification before I +2?10:43
jlibosvaihrachys: I do10:43
jlibosvaihrachys: if it's the ovs compilation patch10:43
ihrachyscool, waiting for the new one10:43
ihrachysyes10:43
jschwarzkevinbenton, it's a good starting point either way10:44
jschwarzkevinbenton, thanks a lot for the help :)10:44
jschwarzkevinbenton, are you still in the US, or in Europe someplace?10:44
kevinbentonjschwarz: US10:45
jschwarzkevinbenton, so a late night again?10:45
kevinbentonjschwarz: yep :)10:45
jschwarzkevinbenton, XD10:45
evrardjpit could have been an early morning!10:45
openstackgerritJakub Libosvar proposed openstack/neutron: gate_hook: Compile ovs only with supported kernels  https://review.openstack.org/36488010:45
*** wolverineav has joined #openstack-neutron10:46
jlibosvaihrachys: ^^10:46
evrardjpkevinbenton: you're working too much :p10:46
kevinbentonevrardjp: nah, was mostly just watching netflix then i remembed i wanted to get an issue figured out before RC1 :)10:47
evrardjpwhen is the rc1 tagged?10:48
evrardjpisn't it one week from now?10:48
*** vishwanathj has joined #openstack-neutron10:48
kevinbentonevrardjp: i think so10:49
kevinbentonrossella_s: so now i'm really confused. i don't know where these microseconds are leaking from :)10:49
kevinbentonrossella_s: https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/_utils.py#L20-L2910:49
kevinbentonrossella_s: that comment is a lie meant to mislead people who read english :)10:50
kevinbentonrossella_s: it definitely chops off all microseconds10:50
rossella_skevinbenton, lol10:50
evrardjptricky text indeed :p10:53
kevinbentonrossella_s: so my guess would actually be that the OVO to_dict method we have isn't calling stringify on the fields10:54
kevinbentonrossella_s: str(datetime.datetime.now())10:55
kevinbenton'2016-09-05 03:54:01.082460'10:55
openstackgerritMerged openstack/neutron-lib: Add eventlet hacking check UT  https://review.openstack.org/36242210:55
openstackgerritMerged openstack/neutron-lib: Fix confusion between Neutron and FWaaS versions  https://review.openstack.org/36274910:55
rossella_skevinbenton, you are probably right, let me check10:56
*** markvoelker has joined #openstack-neutron10:56
evrardjpI think I'm hitting a few bugs with neutron, but I'm not experienced enough to see if that's a feature or not :p It's a long story so bear with me... I've a neutron agent machine that changed hostname.10:58
evrardjpI've disabled the agent with the old hostname, and removed it. The new agent with the new hostname registered, so fine.10:58
*** john-dav_ is now known as john-davidge10:59
evrardjpHowever, I've seen ports whose binding:hostid are still on the old agent (on top of the new ones)10:59
evrardjpoh yes, I forgot to mention that I'm upgrading neutron from liberty to mitaka using openstack-ansible11:00
*** rodrigods has quit IRC11:00
*** rodrigods has joined #openstack-neutron11:00
rossella_skevinbenton, https://github.com/openstack/neutron/blob/master/neutron/objects/base.py#L143-L152 mmm we can to_dict of the base class11:00
*** ababich has joined #openstack-neutron11:00
evrardjplong story short, those ports with the old hostnames are in "BUILD" state11:00
*** markvoelker has quit IRC11:00
evrardjpand even if I put them in admin state down, they will still appear in the dhcp received by my instances11:01
evrardjpSo I think I'm hitting two bugs: port binding host_id does not update when removing agent11:01
jschwarzkevinbenton, can you have a look at https://review.openstack.org/265672 please?11:02
evrardjpdnsmask points to all dhcp ports even if they are in disabled state11:02
jschwarzkevinbenton, those are still happening and it'll be a nice fix for Newton11:02
evrardjpam I wrong?11:02
*** edand has quit IRC11:03
evrardjp(I meant dnsmasq is configured with all dhcp ports for name resolution by default, including disabled ones)11:04
kevinbentonevrardjp: dhcp doesn't care about admin state of ports11:05
kevinbentonevrardjp: so that's expected behavior11:05
*** ijw has joined #openstack-neutron11:05
evrardjpok that's not a bug then, it's just my expectations that were wrong11:05
kevinbentonevrardjp: changing the host name will not automatically update anything11:05
kevinbentonevrardjp: an agent doesn't store state11:06
evrardjpthe agent running with the new hostname will register11:06
evrardjpbut yes, I understand what you mean11:06
kevinbentonevrardjp: right, that binding:host_id field is something that is set by nov11:06
kevinbentonnova*11:06
kevinbentonevrardjp: so unless you manually update it as an admin or via the DB, it's not going to be changed11:06
kevinbentonevrardjp: (assuming you are talking about VM ports here)11:07
evrardjpyup11:07
kevinbentonevrardjp: so the BUILD state will stay because neutron is going to reject status updates from agents about ports they don't own11:07
evrardjpok then I could in my upgrade procedure think about updating the hostname included in the binding:host_id field11:07
kevinbentonevrardjp: yep11:07
*** grauzikas has quit IRC11:08
evrardjpthe problem then would be to put this port outside build state11:08
evrardjpbecause it's technically the same host11:08
evrardjpwell I'll think of a work around11:09
evrardjpthanks for unblocking me already kevinbenton!11:09
kevinbentonevrardjp: i'm not sure what you mean11:09
evrardjpI'll come back later with a plan and discuss it then :)11:09
kevinbentonevrardjp: what do you mean by putting a port outside of the build state?11:09
kevinbentonevrardjp: the agent will update the port to active right when you bring it online after it wires it if the host_id is correct11:10
*** lujinluo has quit IRC11:10
evrardjpyup, so basically after my upgrade, I have for the same "host" different agents (because different hostnames), so I see for my network 2 dhcp agents11:10
evrardjpone port stays in build state (old hostname) and the new one is in status ok11:10
evrardjphere I guess the problem would be the duplication part.11:11
openstackgerritgaryk proposed openstack/neutron: Trunk: remove database transactions from the plugin  https://review.openstack.org/36545911:11
evrardjpI'll think of having a good way to clean this, and I'll come back here11:11
evrardjpI guess it's better11:11
kevinbentonevrardjp: just delete dead agents?11:12
evrardjpthat's already done11:13
evrardjpthere are ports that are still there 'though11:13
*** pc_m has quit IRC11:13
evrardjpI disable and delete the agents with the old hostnames during my procedure11:14
evrardjpI don't delete the dangling ports, which I think I will do11:14
evrardjpit just consumes an IP -- the old dhcp agent has one IP, and the newly registered one will have another ip11:15
*** sleviim1 has joined #openstack-neutron11:15
garykihrachys: i addressed your comments on the objects in transactions11:16
*** sleviim1 has left #openstack-neutron11:16
evrardjpso we'll have an issue for instance's domain name resolution until dhcp has updated the resolv.conf with the new ports only11:17
openstackgerritIhar Hrachyshka proposed openstack/neutron: Introduce ovo objects for ports  https://review.openstack.org/35136811:17
kevinbentonjschwarz: there?11:17
*** rmart04 has quit IRC11:17
kevinbentonevrardjp: you shouldn't have two ports though11:17
kevinbentonevrardjp: for an instance11:17
*** sdake has joined #openstack-neutron11:18
*** amotoki has joined #openstack-neutron11:18
kevinbentonevrardjp: or are you just talking about DHCP ports?11:18
evrardjpmy expectations were that if I disable and delete an old agent, all the dedicated ports assigned to it would be rescheduled or something11:18
evrardjplike dhcp11:18
evrardjphere dhcp ports11:18
*** tflynn has joined #openstack-neutron11:18
*** rmart04 has joined #openstack-neutron11:19
jschwarzkevinbenton, here11:19
*** pc_m has joined #openstack-neutron11:20
kevinbentonevrardjp: IIRC if the dhcp agent is dead, the network will be rescheduled to an active agent11:21
evrardjpBasically I wanted to have the dhcp port for the agent with the new hostname continue to have the same ip -- which won't be as simple as I expected11:21
kevinbentonevrardjp: so is updating the host_id of the port not working then?11:22
evrardjpit's not working11:22
kevinbentonevrardjp: what does it do?11:22
*** lucasagomes is now known as lucas-hungry11:23
evrardjpthe port is correctly updated in the db, and it looks fine at first11:23
*** tflynn has quit IRC11:23
evrardjp(just to make sure I am talking about a dhcp port)11:23
*** sdake has quit IRC11:23
evrardjpbut the status stays in build state11:23
kevinbentonjschwarz: i don't understand the purpose of this patch. why do we go through all of the effort filtering partially formed HA routers on the server side if we have to write the agent to handle receiving them?11:23
*** ushkalim has quit IRC11:23
jschwarzkevinbenton, so the scenario liuyulong|away mentioned earlier makes sense:11:24
evrardjpwhere theoretically it should be ok, because the updated configuration would point to the new hostname, which is the same host, so continuity in the data plane11:24
kevinbentonevrardjp: did you restart the L2 agent?11:24
ihrachyskevinbenton: https://bugzilla.redhat.com/show_bug.cgi?id=1372470 before we throw resources on that one, do you by any chance know about any offending candidates landed in Newton?11:25
openstackbugzilla.redhat.com bug 1372470 in openstack-neutron "Neutron API Performance regression" [High,New] - Assigned to amuller11:25
ihrachyskevinbenton: tl;dr we seem to have api performance slowdown in N11:26
jschwarzkevinbenton, an agent calls sync_routers, while concurrently an API thread deletes the router11:26
evrardjpduring my upgrade process yes, this is why I have 2 dhcp ports registered for the network -- One whose hostid points to my old hostname and one whose hostid points to new hostname11:26
openstackgerritMerged openstack/neutron: Revisit support for trunk segmentation types  https://review.openstack.org/35612711:26
evrardjpbut I update the hostid AFTER the restart of the l2agent11:26
jschwarzkevinbenton, so basically while the agent tries to get the data, the router's interfaces are deleted11:26
kevinbentonihrachys: yes11:26
kevinbentonihrachys: give me a minute11:27
evrardjpfor the port of the old hostname11:27
evrardjpI know that's a mess :p11:27
ihrachyskevinbenton: great. :)11:27
kevinbentonevrardjp: oh, the device_id of the DHCP port is based on the host as well11:28
kevinbentonevrardjp: so that may cause an issue11:28
kevinbentonevrardjp: But the L2 agent would need to be restarted 'after' updating the host_id11:29
kevinbentonevrardjp: anyway because it will not being paying attention to the port at that point11:29
evrardjpok11:30
*** stanzgy has quit IRC11:30
evrardjpI'll have a look at that11:30
*** abregman_ is now known as abregman|mtg11:30
kevinbentonevrardjp: https://github.com/openstack/neutron/blob/master/neutron/common/utils.py#L207-L21311:30
kevinbentonevrardjp: so you would also need to update the device_id to that11:30
kevinbentonevrardjp: that's what the dhcp agent will look for on the server before creating it's own port11:30
evrardjphaha thanks!11:31
evrardjpgreat11:31
kevinbentonjschwarz: ok11:31
kevinbentonjschwarz: that doesnt make sense to me11:31
kevinbentonjschwarz: if the router HA interface is deleted, why is anything returned to the agent11:31
kevinbentonrossella_s: do you have a suggested fix for that?11:33
kevinbentonrossella_s: i'm not sure how dict(self.items()) resolves into things on the fields11:33
kevinbentonrossella_s: maybe the internals of 'dict' call 'str'11:33
kevinbentonihrachys: that bug report has links i can't click!11:34
*** thorst has quit IRC11:34
ihrachyskevinbenton: bad bug report! :(11:35
rossella_skevinbenton, didn't check what the internal of dict does11:35
jschwarzkevinbenton, point!11:35
ihrachyskevinbenton: actually, I was a bit slow on reading my correspondence, I see that the reporter sent an internal email saying that fixing the number of workers helped in his tests; he did not seem to update the bug report yet though; not sure with that in mind I should have dragged you.11:36
rossella_skevinbenton, I'd put a to_string there and see if it fixes the issue...just to check that's the root cause, then we can dig and fix it properly11:36
*** thorst has joined #openstack-neutron11:37
openstackgerritIhar Hrachyshka proposed openstack/neutron-lib: Add support for upper-constraints.txt for all tox targets  https://review.openstack.org/36556011:37
ihrachysjlibosva: am I making sense in https://review.openstack.org/#/c/365459/1//COMMIT_MSG ?11:38
ihrachysnot sure what's our strategy for trunk in process events11:38
kevinbentonihrachys: ack11:38
kevinbentonihrachys: there is a linear query increase hiding in the segments plugin11:38
kevinbentonihrachys: but that will only impact you if you have it enabled11:39
*** ushkalim has joined #openstack-neutron11:39
kevinbentonrossella_s: ok. i'll dig into it more in a bit. i can't run code on the current computer i'm on11:40
kevinbentonrossella_s: (other than the simple python REPL stuff)11:40
kevinbentonrossella_s: will try once i get an env with all of the dev stuff on it11:40
rossella_skevinbenton, ack11:41
*** liuyulong_ has joined #openstack-neutron11:41
*** thorst has quit IRC11:41
ihrachyskevinbenton: ack. thanks for the info.11:43
openstackgerritIhar Hrachyshka proposed openstack/neutron: Introduce ovo objects for networks  https://review.openstack.org/26965811:43
*** gongysh has joined #openstack-neutron11:44
*** liuyulong|away has quit IRC11:45
jschwarzliuyulong_, so11:45
jschwarzliuyulong_, I discussed this with kevinbenton just a few moments ago11:45
jschwarzkevinbenton, the logic that deals with "this router doesn't have HA interfaces" was removed IMO from the code when we introduced the allocating logic11:46
*** amotoki has quit IRC11:46
jschwarzkevinbenton, let me find the patch that did it11:47
jschwarzkevinbenton, https://review.openstack.org/#/c/257059/40/neutron/db/l3_hamode_db.py@70811:48
jschwarzkevinbenton, so basically we can consider this a regression? :P and quite an easy fix IMO11:48
*** gongysh has quit IRC11:49
jlibosvaihrachys: I'm afraid I can't tell. I don't know why those notifications are there11:49
*** salv-orlando has quit IRC11:51
*** salv-orlando has joined #openstack-neutron11:51
kevinbentonjschwarz: yes, we assumed that ALLOCATING was going to filter for us11:52
kevinbentonjschwarz: which was wrong11:52
kevinbentonjschwarz: put it back and call it a day11:52
jschwarzkevinbenton, yippe11:52
jschwarzkevinbenton, rc1 material?11:52
kevinbentonjschwarz: yes, i believe so11:52
jschwarzkevinbenton, I'll handle it11:53
*** tbachman has joined #openstack-neutron11:54
ihrachysjlibosva: remember at midcycle we were comparing approaches of callbacks vs. ext manager? that seems to be a relevant discussion. If we assume payload events part of API to implement trunks in plugins, then it would make sense to maintain events that may be unneeded for ml2 for the sake of other consumers.11:55
*** markvoelker has joined #openstack-neutron11:57
jlibosvaihrachys: I agree. As long as we have other consumers11:57
jschwarzkevinbenton, do you understand the comment here: https://github.com/openstack/neutron/blob/master/neutron/db/l3_hamode_db.py#L718 ?11:58
jschwarzkevinbenton, (this is where the filter should be going back in)11:58
*** abregman|mtg has quit IRC12:00
kevinbentonjschwarz: https://github.com/openstack/neutron/commit/179b8301edad50f999417f52b77092a496fb448e12:00
kevinbentonjschwarz: so it sounds like it should already be skipped12:00
*** ijw has quit IRC12:01
*** dkehn_ has quit IRC12:01
*** dkehn has quit IRC12:01
jschwarzkevinbenton, aye, was just looking at it12:01
*** markvoelker has quit IRC12:01
jschwarzkevinbenton, so how can a L3HARouterAgentPortBinding not have a binding.port?12:02
jschwarzkevinbenton, it's always created with one (and only the l3_agent_id is added afterwards)12:02
kevinbentonjschwarz: is it nullable?12:02
jschwarzkevinbenton, nope12:03
jschwarzkevinbenton, nullable=False12:03
kevinbentonjschwarz: so where do you see one without a port?12:03
jschwarzkevinbenton, https://review.openstack.org/#/c/265680/22/neutron/db/l3_hamode_db.py@70212:03
kevinbentonjschwarz: oh, well if the port is deleted after you have the binding, the relationship lookup isn't going to return a port12:04
kevinbentonjschwarz: that relationship is lazily loaded12:04
jschwarzkevinbenton, :<12:04
*** davidsha has quit IRC12:06
*** ushkalim has quit IRC12:06
*** davidsha has joined #openstack-neutron12:07
*** dkehn has joined #openstack-neutron12:08
*** qwebirc72994 has joined #openstack-neutron12:09
jschwarzkevinbenton, I'm not clear as to how this issue is popping up exactly12:09
*** fzdarsky has quit IRC12:09
jschwarzkevinbenton, I'll have a few runs on my local setup and find the fault12:09
*** annp has quit IRC12:10
*** prateek_ has joined #openstack-neutron12:11
*** ilyashakhat_mobi has joined #openstack-neutron12:11
*** prateek has quit IRC12:11
*** fzdarsky has joined #openstack-neutron12:12
*** abregman|mtg has joined #openstack-neutron12:12
*** ilyashakhat_mobi has quit IRC12:13
*** apuimedo is now known as apuimedo|lunch12:16
*** parora has joined #openstack-neutron12:17
kevinbentonjschwarz: ack12:19
*** ushkalim has joined #openstack-neutron12:20
*** dkehn_ has joined #openstack-neutron12:21
*** prateek has joined #openstack-neutron12:21
*** prateek_ has quit IRC12:21
*** parora has quit IRC12:22
*** marksei has joined #openstack-neutron12:22
*** jpena is now known as jpena|lunch12:23
*** lucas-hungry is now known as lucasagomes12:25
*** amotoki has joined #openstack-neutron12:27
*** jckasper has joined #openstack-neutron12:28
*** itzikb has joined #openstack-neutron12:29
slaweq_jlibosva: hello, I wrote one rfe: https://bugs.launchpad.net/neutron/+bug/162027912:29
openstackLaunchpad bug 1620279 in neutron "Allow metadata agent to make calls to more than one nova_metadata_ip" [Undecided,New]12:29
slaweq_jlibosva: can You maybe have a look on that and tell me if it makes sense for You? thx in advance12:29
jlibosvaslaweq_: hi12:30
jlibosvaslaweq_: looking12:30
jlibosvaslaweq_: if there are more nova apis in the cluster, won't they be hidden behind a VIP?12:30
*** gongysh has joined #openstack-neutron12:31
slaweq_jlibosva: I consider case when one of nova-api will return for example error 500 - then haproxy (or other LB) will still pass it to agent and agent to instance12:31
*** Guest64781 is now known as zeus12:31
*** zeus has joined #openstack-neutron12:32
jlibosvaslaweq_: from my experience cloud-init tries mutliple times if it doesn't succeed12:32
*** jckasper has quit IRC12:32
slaweq_but it can be that other nova services are working fine so repeat request to another nova api could return proper data and then user on instance will not even see that something is wrong with metadata server12:32
jlibosvaslaweq_: I see12:33
*** ilyashakhat_mobi has joined #openstack-neutron12:33
*** qwebirc72994 has quit IRC12:33
slaweq_jlibosva: yes, but haproxy can pass it to node with broken nova-api (it might be not dead but still broken)12:33
*** apuimedo|lunch is now known as apuimedo12:38
*** wolverineav has quit IRC12:38
jlibosvaslaweq_: I'm just thinking whether this will introduce a complexity to mtd agent while it can be solved by having round-robin LB on nova api level12:41
jlibosvaslaweq_: I'm not saying it's wrong thing to do :) Just thinking out loud12:42
*** yedongcan has left #openstack-neutron12:44
slaweq_jlibosva: of course :) I think it's not big change in code and it can be useful for example for public cloud operators which has got external customers (like OVH) because customers don't like to have "random" error 500 while instance is booting12:44
*** sleviim has quit IRC12:44
jlibosvaslaweq_: yeah, the question is whether it could be caused by wrong cluster configuration. In the guest, you can get 500 once and then succeed while not even noticing as everything goes smooth afterwards. But it sounds like you have more experience with this failure than I do :)12:45
slaweq_jlibosva: we had something like that in some cases, maybe it wasn't very common situation but we had such12:50
slaweq_and we made similar patch internally which works fine for us, so I want to propose it for upstream now :)12:51
*** lilintan has quit IRC12:52
jlibosvaslaweq_: have you tried to have a loadbalancer for nova-api and just retry from metadata agent to the same VIP? It sounds way easier to me with much smaller change12:54
slaweq_jlibosva: I don't remember exactly but I think that we tried and made such patch because of reason which I don't remember now :/12:56
*** ilyashakhat_mobi has quit IRC12:56
*** garyk has quit IRC12:59
jschwarzkevinbenton, got it. looks like an easy fix. I'll add you as a reviewer.12:59
openstackgerritMerged openstack/neutron: QOS-DB: remove deprecation warning  https://review.openstack.org/36550313:01
kevinbentonjschwarz: ack13:01
openstackgerritIhar Hrachyshka proposed openstack/neutron: objects: expose database model for NeutronDbObject instances  https://review.openstack.org/34827913:02
openstackgerritIhar Hrachyshka proposed openstack/neutron: objects: remove support for multiple db models in from_db_object  https://review.openstack.org/34827113:02
ihrachyselectrocucaracha: hi13:04
*** EinstCrazy has joined #openstack-neutron13:04
ihrachyselectrocucaracha: how is it going with the patch to avoid raising InvalidInput on unknown query filters from api layer?13:05
*** abregman|mtg has quit IRC13:09
kevinbentonrossella_s: hmm. we're in a bit of a weird spot13:11
kevinbentonrossella_s: what is to_dict() intended to be used for, just for serialization to the API?13:11
kevinbentonrossella_s: or will it be passed around for other things to use13:12
kevinbentonrossella_s: because right now to_dict() is iterating over self13:12
kevinbentonrossella_s: which provides access to the native datetime object for the timestamp fields13:12
*** yedongcan has joined #openstack-neutron13:13
openstackgerritGenadi Chereshnya proposed openstack/neutron: Checking functionality of DVR  https://review.openstack.org/34038313:13
kevinbentonrossella_s: and then when str is called much later during conversion for the API, it's way too late13:13
kevinbentonrossella_s: so the question is, do we want to_dict() to call stringify right then?13:13
kevinbentonrossella_s: because if we do, then it's a little less useful if other things wanted to use things other than strings13:14
dasmihrachys: there is a Labor Day in US, so electrocucaracha is unavailable.13:17
ihrachysdasm: that's insane!13:17
ihrachysdasm: how is it a labour day if no one works?13:17
dasmihrachys: I know, I know.. :D13:18
* ihrachys shakes his head13:18
jlibosvadasm: you still in Europe so you're forced to work? :-P13:18
*** rossella_s has quit IRC13:18
dasmjlibosva: shh... no one noticed it... yet :P13:18
*** rossella_s has joined #openstack-neutron13:19
dasmihrachys: we can think about calling all other days as non-Labor Day, and force everyone to work.13:19
electrocucarachaThanks dasm13:19
dasmelectrocucaracha: oh... you're alive! :D13:19
ihrachysnow that's the spirit guys13:19
*** oanson has quit IRC13:20
ihrachyselectrocucaracha: ok, just quick question: do you work on the bit, or I can respin myself?13:20
dasmihrachys: he probably just want to say "Hi" and disappear.13:20
electrocucarachaihrachys: I'm going to submit what I have but it seems to simple to be truth13:20
ihrachyselectrocucaracha: ok, I will wait your bit then13:20
ihrachyselectrocucaracha: enjoy the no-labour labour day13:21
electrocucarachaihrachys:  actually I was planning to attend the upgrades meeting13:21
*** abregman|mtg has joined #openstack-neutron13:22
electrocucarachaihrachys: specifically because there are two things that the team has faced13:22
electrocucarachaihrachys: one is the outer joins13:23
electrocucarachaihrachys: and the last one is regarding that some methods are passing the session instead of the db context13:24
electrocucarachaihrachys: but let's wait until the meeting13:24
* electrocucaracha goes for breakfast13:25
*** oshvartz has quit IRC13:25
*** marksei has quit IRC13:25
*** oshvartz has joined #openstack-neutron13:25
*** nyechiel has quit IRC13:25
*** marksei has joined #openstack-neutron13:25
*** ushkalim has quit IRC13:25
*** itzkb_ has joined #openstack-neutron13:27
*** itzikb has quit IRC13:27
*** baoli has joined #openstack-neutron13:28
rossella_skevinbenton, to_dict might be passed around for other things to use so I wouldn't call stringify right then13:29
rossella_skevinbenton, we could override to_dict for datetime object sothat it calls stringify?13:31
rossella_sihrachys, what do you think?13:31
kevinbentonrossella_s: well i was just thinking to_dict() is supposed to be the equivalent of make_port_dict, etc right?13:31
*** jpena|lunch is now known as jpena13:31
ihrachyskevinbenton: it's not supposed to apply extensions13:32
kevinbentonihrachys: not from extensions context13:32
kevinbentonihrachys: from format context13:32
* ihrachys notes he misses the context13:32
kevinbentonihrachys: is it reasonable to say that everything should be strings?13:32
ihrachyskevinbenton: where? In to_dict result?13:32
kevinbentonihrachys: right now created_at/updated_at are putting datetime.datetime objects in there13:32
kevinbentonihrachys: yep13:32
rossella_skevinbenton, we can enforce that13:33
kevinbentonihrachys: which leaks microseconds eventually when they get str called to them13:33
*** baoli_ has joined #openstack-neutron13:33
kevinbentonrossella_s: so i'm working on a patch to use obj_to_primitive() instead of dict(self.items())13:33
*** jprovazn has joined #openstack-neutron13:33
rossella_skevinbenton, ack13:33
kevinbentonto ensure everything is converted into its appropriate serialized format13:34
*** abregman|mtg is now known as abregman13:36
*** baoli has quit IRC13:36
ihrachyskevinbenton: modifying the base class to_dict?13:36
*** john-davidge has quit IRC13:37
*** ushkalim has joined #openstack-neutron13:37
*** john-davidge has joined #openstack-neutron13:37
*** links has quit IRC13:38
ihrachyskevinbenton: till now, it's not necessarily strings, booleans or integers are fine too13:39
jschwarzkevinbenton, wait for it...13:39
openstackgerritJohn Schwarz proposed openstack/neutron: l3 ha: don't send routers without '_ha_interface'  https://review.openstack.org/36565313:39
jschwarzkevinbenton, legendary :P13:39
*** ababich has quit IRC13:40
*** gongysh has quit IRC13:42
*** gongysh has joined #openstack-neutron13:44
*** irenab has quit IRC13:46
*** thorst has joined #openstack-neutron13:46
*** jlanoux has joined #openstack-neutron13:47
*** ilyashakhat_mobi has joined #openstack-neutron13:52
*** luis5tb has quit IRC13:52
*** jhershbe_ has quit IRC13:52
openstackgerritVictor Morales proposed openstack/neutron: Make optional the validation of filters  https://review.openstack.org/36565913:53
*** jhershbe has quit IRC13:53
*** EinstCrazy has quit IRC13:54
*** wolverineav has joined #openstack-neutron13:55
*** baoli_ has quit IRC13:56
*** baoli has joined #openstack-neutron13:56
*** markvoelker has joined #openstack-neutron13:57
*** amotoki has quit IRC13:58
*** ilyashakhat_mobi has quit IRC14:00
*** jistr is now known as jistr|afk14:01
*** ushkalim has quit IRC14:01
*** markvoelker has quit IRC14:03
*** lihi has quit IRC14:03
*** hynekm has quit IRC14:04
kevinbentonjschwarz: that comment implies there was a problem there though with DVR+HA routers14:06
*** thorst has quit IRC14:06
kevinbentonjschwarz: did you figure that out?14:06
jschwarzkevinbenton, negative14:06
jschwarzkevinbenton, I looked back at the exact patchset where it was added and there weren't any comments regarding this issue14:06
*** thorst has joined #openstack-neutron14:07
kevinbentonjschwarz: added by liuyulong_, right?14:07
jschwarzkevinbenton, yes, but I think he's gone to bed14:07
jschwarzkevinbenton, also that comment doesn't make a lot of sense14:07
jschwarzkevinbenton, if it's a DVR+HA in dvr_snat, it should have this attribute (as it's an HA router that is scheduled to that agent)14:08
kevinbentonjschwarz: well we should def wait for liuyulong_ to comment. seems like it's explicit about breaking sync of DVR+HA on the compute node14:08
jschwarzkevinbenton, and if it's a dvr agent then it shouldn't get there14:08
kevinbentonjschwarz: the dvr agent syncs the router14:08
jschwarzkevinbenton, actually I have a setup with that right now14:08
jschwarzlet me have a looksie14:08
*** claudiub|2 has joined #openstack-neutron14:09
*** pradk has joined #openstack-neutron14:09
*** prateek_ has joined #openstack-neutron14:11
*** claudiub has quit IRC14:12
*** wolverineav has quit IRC14:13
*** irenab has joined #openstack-neutron14:13
*** prateek has quit IRC14:14
*** wolverineav has joined #openstack-neutron14:14
*** ushkalim has joined #openstack-neutron14:14
*** thorst has quit IRC14:15
*** marksei has quit IRC14:15
*** mihero has quit IRC14:16
*** liuyulong has joined #openstack-neutron14:16
jschwarzkevinbenton, ah yes14:17
jschwarzkevinbenton, I'll fix the condition14:17
liuyulonghi there14:18
*** gongysh has quit IRC14:18
*** Alex_Stef has quit IRC14:18
*** ekuris has quit IRC14:19
*** EinstCrazy has joined #openstack-neutron14:19
*** EinstCrazy has quit IRC14:19
*** parora has joined #openstack-neutron14:19
*** prateek_ has quit IRC14:22
jschwarzliuyulong, have a look at this:14:25
jschwarzliuyulong, https://review.openstack.org/#/c/365653/114:25
*** john-davidge has quit IRC14:26
*** jckasper has joined #openstack-neutron14:27
*** jckasper has quit IRC14:28
liuyulongjschwarz, hi, looking at it14:28
*** buttercup has joined #openstack-neutron14:29
*** jistr|afk is now known as jistr14:29
*** salv-orl_ has joined #openstack-neutron14:30
openstackgerritgaryk proposed openstack/neutron-dynamic-routing: Use model_base from neutron-lib  https://review.openstack.org/36567914:30
*** salv-orlando has quit IRC14:32
*** ushkalim has quit IRC14:32
*** chlong has joined #openstack-neutron14:33
openstackgerritJakub Libosvar proposed openstack/neutron: gate_hook: Compile ovs only with supported kernels  https://review.openstack.org/36488014:34
jlibosvaihrachys: ^^ this time hopefully not a screwed up patch, thanks moshele :)14:34
moshelejlibosva: np :)14:35
dasmjlibosva: but you still didn't update commit message :(14:36
jlibosvadasm: yeah, that's because I'm a liar14:36
*** irenab has quit IRC14:36
dasm:D14:37
*** irenab has joined #openstack-neutron14:37
*** armax has joined #openstack-neutron14:38
kevinbentonihrachys, rossella_s: it's time for me to get some breakfast. it would be cool if you can look at ^^14:38
*** jhershbe_ has joined #openstack-neutron14:39
kevinbenton(i suppose i should have uploaded a patch before saying that)14:39
ihrachyskevinbenton: not seeing ^^14:39
kevinbentonTHIS IS THE MOST EMBARASSING DAY OF MY LIFE14:39
ihrachyskevinbenton: that's ok, enjoy your late dinner :P14:39
openstackgerritKevin Benton proposed openstack/neutron: Convert OVO fields to primitives for to_dict()  https://review.openstack.org/36568414:39
kevinbentonihrachys, rossella_s: ^^14:39
kevinbentonthat one :)14:39
openstackgerritJakub Libosvar proposed openstack/neutron: gate_hook: Compile ovs only with supported kernels  https://review.openstack.org/36488014:39
ihrachyskevinbenton: aye. please move away from the kbd, it sucks your sole14:40
*** armax has quit IRC14:40
*** irenab has quit IRC14:40
kevinbentonihrachys: we have another slight issue14:41
kevinbentonihrachys: i don't think our current format puts on TZ14:41
*** jhershbe has joined #openstack-neutron14:41
kevinbentonihrachys: but OVO does14:41
kevinbentonihrachys: https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/_utils.py#L24-L2914:41
ihrachyskevinbenton: custom type needed?14:41
*** irenab has joined #openstack-neutron14:41
*** nyechiel has joined #openstack-neutron14:42
kevinbentonihrachys: https://github.com/openstack/neutron/blob/master/neutron/services/timestamp/timestamp_db.py#L95-L9914:42
kevinbentonihrachys: either that or we 'fix' our current one to include that info14:42
kevinbentonihrachys: unfortunately it could be seen as a breaking API change14:43
kevinbentonihrachys: but we would be consistent with nova at that point at least14:43
jlibosvakevinbenton: looking briefly on that patch - it seems like we're approaching to self.obj_to_primitive() - would it make sense to just return dehydrated object to API since we're now introducing more complex object fields?14:43
kevinbentonihrachys: this actually came up before https://bugs.launchpad.net/neutron/+bug/156120014:44
openstackLaunchpad bug 1561200 in neutron "created_at and updated_at times don't include timezone" [Undecided,Expired]14:44
kevinbentonihrachys: would be interesting to get your thoughts on that14:44
ivasilevskayajlibosva, a quick question about ovsfw and ubuntu14.04. Am I right that for 3.13 kernel I need only openvswitch-common and openvswitch-switch packages?14:44
kevinbentonjlibosva: obj_to_primitive() doesn't appear to include some of the extra fields (e.g. tenant_id). that was what I first started with self.obj_to_primitive()['verioned_object.data']14:45
jlibosvaivasilevskaya: as per https://github.com/openvswitch/ovs/blob/branch-2.5/FAQ.md#q-are-all-features-available-with-all-datapaths you'd need to compile openvswitch kernel module in order to introduce conntrack datapath14:46
*** ushkalim has joined #openstack-neutron14:46
ihrachyskevinbenton: I will check it later today or tomorrow. I hope it's ok.14:47
kevinbentonihrachys: ok. just needs to be fixed before release14:47
ihrachysack14:47
*** itzkb_ has quit IRC14:49
ivasilevskayajlibosvar, thanks. So not having vm connectivity without this step is an expected behavior, right?14:49
*** moshele has quit IRC14:50
jlibosvaivasilevskaya: I'd expect that ovs-ofctl commands will fail printing some binary characters to ovs-agent logs. yeah, i think it'll lead to no connectivity on vms14:50
jlibosvaivasilevskaya: you can use sanity checks to test whether your box supports conntrack with ovs14:50
jlibosvaivasilevskaya: neutron-sanity-check --ovs_conntrack14:51
openstackgerritKevin Benton proposed openstack/neutron: Convert OVO fields to primitives for to_dict()  https://review.openstack.org/36568414:51
openstackgerritKevin Benton proposed openstack/neutron: Add common way to extend standard attribute models  https://review.openstack.org/36468114:51
ivasilevskayajlibosva, I owe you a drink. How could I forget to run sanity-check?14:53
ivasilevskayathanks14:53
openstackgerritLIU Yulong proposed openstack/neutron: Resolve infinite loops in L3 HA agent side  https://review.openstack.org/26567214:54
jlibosvakevinbenton: aha, I see. cause we don't have some stuff in fields and obj_to_primitive() iterates over self.fields.items() while we need self.items(). That makes me think whether we should rather introduce a hook in fields somehow instead of re-implementing obj_to_primitive14:54
jlibosvaivasilevskaya: I accept beers as bribes :)14:54
openstackgerritJohn Schwarz proposed openstack/neutron: l3 ha: don't send routers without '_ha_interface'  https://review.openstack.org/36565314:55
jschwarzkevinbenton, ^14:55
jschwarzkevinbenton, took extra attention to the 'dvr' case14:55
openstackgerritIhar Hrachyshka proposed openstack/neutron: objects: expose database model for NeutronDbObject instances  https://review.openstack.org/34827914:58
*** chlong has quit IRC14:58
jschwarzivasilevskaya, it's a trap14:58
jschwarzivasilevskaya, jlibosva also owed me a drink but I never got it when we last met14:58
jlibosvajschwarz: did I?14:59
jschwarzjlibosva, you sure did14:59
jschwarzjlibosva, let me just come up with <generic reason 1337>14:59
ihrachysupgrades meeting starting in #openstack-meeting-alt in 1 mins14:59
jschwarzjlibosva, you broke my code severely and added '1337' in all the routers' names, so you said you'd repay me with a bear if I look the other way15:00
jschwarzjlibosva, specifically a bear and not a beer, take note.15:00
*** nmagnezi_ has quit IRC15:00
*** irenab has quit IRC15:01
jlibosvajschwarz: https://media2.giphy.com/media/glmRyiSI3v5E4/200.gif15:02
openstackgerritgaryk proposed openstack/neutron-dynamic-routing: Use model_base from neutron-lib  https://review.openstack.org/36567915:02
jschwarzjlibosva, that's the second time I've seen this one today15:02
jlibosvajschwarz: I know15:02
*** jb__ has quit IRC15:03
*** thorst has joined #openstack-neutron15:12
liuyulongakamyshnikova, ping15:13
*** saggi has quit IRC15:14
*** baoli has quit IRC15:19
*** thorst has quit IRC15:20
*** liuyulong is now known as liuyulong_zzz15:21
*** jhershbe has quit IRC15:22
*** jhershbe_ has quit IRC15:22
*** oshvartz has quit IRC15:25
*** armax has joined #openstack-neutron15:26
*** vishwanathj has quit IRC15:26
*** liuyulong_zzz has quit IRC15:26
*** devvesa has quit IRC15:28
*** yfried has quit IRC15:31
*** wolverineav has quit IRC15:32
*** pcaruana has quit IRC15:36
*** yedongcan has quit IRC15:42
*** devvesa has joined #openstack-neutron15:42
*** john51_ has joined #openstack-neutron15:44
openstackgerritSlawek Kaplonski proposed openstack/neutron-lib: Add info about API method to get default quotas  https://review.openstack.org/35834415:45
*** nyechiel_ has joined #openstack-neutron15:46
*** inara` has joined #openstack-neutron15:46
*** mirrorbo1 has joined #openstack-neutron15:46
*** rmcadams_ has joined #openstack-neutron15:47
*** kozhukalov_ has joined #openstack-neutron15:47
*** patrickeast_ has joined #openstack-neutron15:47
*** boris-42_ has joined #openstack-neutron15:47
*** serverascode_ has joined #openstack-neutron15:47
*** csatari_ has joined #openstack-neutron15:47
*** carl_baldwin_ has joined #openstack-neutron15:47
*** obondarev has joined #openstack-neutron15:48
*** diltram_ has joined #openstack-neutron15:48
*** marlinc_ has joined #openstack-neutron15:48
*** sc68cal_ has joined #openstack-neutron15:48
*** buttercup has quit IRC15:48
*** rmcadams has quit IRC15:48
*** serverascode has quit IRC15:48
*** flaviof has quit IRC15:48
*** mtreinish has quit IRC15:48
*** briancurtin has quit IRC15:48
*** melwitt has quit IRC15:48
*** sweston has quit IRC15:48
*** boris-42 has quit IRC15:48
*** diltram has quit IRC15:48
*** inara has quit IRC15:48
*** electrocucaracha has quit IRC15:48
*** marlinc has quit IRC15:48
*** cargonza has quit IRC15:49
*** sc68cal has quit IRC15:49
*** barclaac has quit IRC15:49
*** patrickeast has quit IRC15:49
*** zhenguo has quit IRC15:49
*** kozhukalov has quit IRC15:49
*** devvesa has quit IRC15:49
*** nyechiel has quit IRC15:49
*** bkopilov has quit IRC15:49
*** john51 has quit IRC15:49
*** carl_baldwin has quit IRC15:49
*** quade has quit IRC15:49
*** dasanind_ has quit IRC15:49
*** csatari has quit IRC15:49
*** jistr has quit IRC15:49
*** jpena has quit IRC15:49
*** harlowja_ has quit IRC15:49
*** mirrorbox has quit IRC15:49
*** mtreinish has joined #openstack-neutron15:49
*** jpena has joined #openstack-neutron15:49
*** jistr has joined #openstack-neutron15:49
*** shihanzhang has quit IRC15:49
*** devvesa has joined #openstack-neutron15:49
*** bkopilov has joined #openstack-neutron15:49
*** marlinc_ is now known as marlinc15:50
*** shihanzhang has joined #openstack-neutron15:50
*** rmcadams_ is now known as rmcadams15:50
*** carl_baldwin_ is now known as carl_baldwin15:50
*** andreas_s has quit IRC15:50
*** patrickeast_ is now known as patrickeast15:51
*** barclaac has joined #openstack-neutron15:51
*** melwitt has joined #openstack-neutron15:51
*** csatari_ is now known as csatari15:51
*** yedongcan has joined #openstack-neutron15:52
*** kozhukalov_ is now known as kozhukalov15:52
*** flaviof has joined #openstack-neutron15:52
*** melwitt is now known as Guest7823615:52
*** SKipp_ has quit IRC15:53
*** vishwanathj has joined #openstack-neutron15:54
*** vishwanathj has quit IRC15:54
*** boris-42_ is now known as boris-4215:55
*** buttercup has joined #openstack-neutron15:56
*** serverascode_ is now known as serverascode15:57
*** markvoelker has joined #openstack-neutron15:58
*** dasanind_ has joined #openstack-neutron15:58
*** sweston has joined #openstack-neutron15:58
*** marun has quit IRC16:00
*** quade has joined #openstack-neutron16:01
*** prateek_ has joined #openstack-neutron16:01
*** SKipp_ has joined #openstack-neutron16:02
*** rmart04 has quit IRC16:02
*** cargonza has joined #openstack-neutron16:02
*** xdcc has joined #openstack-neutron16:02
*** markvoelker has quit IRC16:03
*** matrohon has quit IRC16:03
*** ushkalim has quit IRC16:04
*** briancurtin has joined #openstack-neutron16:04
*** parora has quit IRC16:04
*** zhenguo has joined #openstack-neutron16:05
*** yedongcan has quit IRC16:05
openstackgerritYushiro FURUKAWA proposed openstack/python-neutronclient: Add FWaaS V2 commands for OSC plugin  https://review.openstack.org/35158216:05
openstackgerritYushiro FURUKAWA proposed openstack/python-neutronclient: Add common utilities for OSC plugin implementation  https://review.openstack.org/34809716:05
*** obondarev has quit IRC16:06
*** jlanoux has quit IRC16:07
*** bfernando has quit IRC16:10
*** armax has quit IRC16:11
*** itisha has joined #openstack-neutron16:16
*** thorst has joined #openstack-neutron16:18
*** thorst has quit IRC16:26
*** abregman has quit IRC16:28
*** jckasper has joined #openstack-neutron16:28
*** tmorin has quit IRC16:29
*** jckasper has quit IRC16:33
openstackgerritMerged openstack/neutron: Check MTU sanity of trunk port subports  https://review.openstack.org/36291716:36
*** mickeys has joined #openstack-neutron16:36
*** devvesa has quit IRC16:37
*** jlibosva has quit IRC16:38
openstackgerritJakub Libosvar proposed openstack/neutron: OVSDBHandler for trunk ports  https://review.openstack.org/34746616:38
*** rmart04 has joined #openstack-neutron16:40
*** thorst has joined #openstack-neutron16:44
*** tesseract- has quit IRC16:45
*** hynekm has joined #openstack-neutron16:48
*** thorst has quit IRC16:49
*** mickeys has quit IRC16:49
*** thorst has joined #openstack-neutron16:50
*** prometheanfire has joined #openstack-neutron16:50
*** davidsha has quit IRC16:50
*** lucasagomes is now known as lucas-dinner16:50
prometheanfireare https://bugs.launchpad.net/neutron/+bug/1505316 and https://bugs.launchpad.net/neutron/+bug/1570122 targeted for newton?16:52
openstackLaunchpad bug 1505316 in neutron "compute port lose fixed_ips on restart l3-agent if subnet is prefix delegated" [Medium,In progress] - Assigned to venkata anil (anil-venkata)16:52
openstackLaunchpad bug 1570122 in neutron "ipv6 prefix delegated subnets are not accessable external of the router they are attached." [Medium,Confirmed]16:52
prometheanfireit's been about one whole release with it broken16:52
*** saggi has joined #openstack-neutron16:53
*** neiljerram has joined #openstack-neutron16:58
*** john-davidge has joined #openstack-neutron16:58
*** thorst has quit IRC16:58
*** neiljerram has quit IRC17:01
*** ranjithd has quit IRC17:02
openstackgerritIhar Hrachyshka proposed openstack/neutron: tests: remove self.db_obj from objects tests  https://review.openstack.org/36575217:04
prometheanfireihrachys: fancy seeing you here17:07
ihrachysprometheanfire: ?17:07
*** Alex_Stef has joined #openstack-neutron17:07
*** akamyshnikova_ has joined #openstack-neutron17:10
*** ihrachys has quit IRC17:10
*** akamyshnikova_ has quit IRC17:13
*** rmart04 has quit IRC17:14
*** rmart04 has joined #openstack-neutron17:14
*** saggi has quit IRC17:15
*** gvrangan has joined #openstack-neutron17:19
*** yamahata has quit IRC17:22
*** armax has joined #openstack-neutron17:24
*** armax has quit IRC17:24
*** slaweq has quit IRC17:25
*** armax has joined #openstack-neutron17:26
*** mosulica has quit IRC17:26
*** armax has quit IRC17:27
*** fzdarsky is now known as fzdarsky|afk17:28
*** oanson has joined #openstack-neutron17:28
*** armax has joined #openstack-neutron17:28
*** armax has quit IRC17:28
*** armax has joined #openstack-neutron17:30
*** nyechiel_ has quit IRC17:30
*** armax has quit IRC17:31
*** eezhova has quit IRC17:32
*** ivc_ has quit IRC17:32
*** donghao has quit IRC17:32
*** armax has joined #openstack-neutron17:34
*** armax has quit IRC17:34
*** mihero has joined #openstack-neutron17:35
*** armax has joined #openstack-neutron17:36
*** armax has quit IRC17:36
*** armax has joined #openstack-neutron17:37
openstackgerritDarek Smigiel proposed openstack/neutron: Make optional the validation of filters  https://review.openstack.org/36565917:38
*** gopenshaw has joined #openstack-neutron17:39
*** jpena is now known as jpena|away17:39
*** hynekm has quit IRC17:40
*** armax has quit IRC17:42
*** saggi has joined #openstack-neutron17:42
openstackgerritDarek Smigiel proposed openstack/neutron: Make optional the validation of filters  https://review.openstack.org/36565917:44
*** marun has joined #openstack-neutron17:45
*** bfernando has joined #openstack-neutron17:45
*** hynekm has joined #openstack-neutron17:46
openstackgerritAnindita Das proposed openstack/neutron: Relocate RouterL3AgentBinding DB model  https://review.openstack.org/35571917:48
openstackgerritHynek Mlnarik proposed openstack/neutron: API tests: Check MTU sanity of trunk/subport  https://review.openstack.org/36420817:49
openstackgerritTerry Wilson proposed openstack/neutron: Handle uuid references within an ovsdb transaction  https://review.openstack.org/35810117:50
*** sambetts is now known as sambetts|afk17:52
*** jpena|away is now known as jpena|off17:52
*** bfernando has quit IRC17:52
*** thorst has joined #openstack-neutron17:56
*** jprovazn has quit IRC17:57
*** markvoelker has joined #openstack-neutron17:59
*** john-davidge has quit IRC17:59
*** parora has joined #openstack-neutron18:02
*** rmart04 has quit IRC18:03
*** thorst has quit IRC18:03
*** markvoelker has quit IRC18:03
*** openstackgerrit has quit IRC18:03
*** openstackgerrit has joined #openstack-neutron18:04
*** prateek has joined #openstack-neutron18:05
*** prateek_ has quit IRC18:05
*** parora has quit IRC18:07
*** nplanel has joined #openstack-neutron18:12
*** prateek has quit IRC18:17
*** nplanel has quit IRC18:20
*** nmagnezi has joined #openstack-neutron18:21
*** john-davidge has joined #openstack-neutron18:22
*** jckasper has joined #openstack-neutron18:24
*** john-davidge has quit IRC18:27
*** pcaruana has joined #openstack-neutron18:31
*** mihero has quit IRC18:31
*** mihero has joined #openstack-neutron18:32
*** rmart04 has joined #openstack-neutron18:41
*** mihero has quit IRC18:41
*** nmagnezi has quit IRC18:42
*** buttercup has quit IRC18:49
*** rmart04 has quit IRC18:49
*** moshele has joined #openstack-neutron18:49
*** hynekm has quit IRC18:50
*** rmart04 has joined #openstack-neutron18:50
*** slaweq has joined #openstack-neutron18:57
*** john-davidge has joined #openstack-neutron18:58
*** eezhova has joined #openstack-neutron18:59
*** thorst has joined #openstack-neutron19:00
*** jlibosva has joined #openstack-neutron19:07
*** thorst has quit IRC19:08
*** oanson has quit IRC19:12
openstackgerritMohit Malik proposed openstack/neutron: Add OVO for AutoAllocateToplogy  https://review.openstack.org/35750619:17
*** armax has joined #openstack-neutron19:19
*** armax has quit IRC19:19
*** mihero has joined #openstack-neutron19:20
*** armax has joined #openstack-neutron19:21
*** rmart04 has quit IRC19:21
openstackgerritShashank Kumar Shankar proposed openstack/neutron: OVO for Provisioning Block DB Model  https://review.openstack.org/35748819:23
*** armax has quit IRC19:25
*** jckasper has quit IRC19:25
*** jckasper has joined #openstack-neutron19:26
*** armax has joined #openstack-neutron19:27
*** jckasper has quit IRC19:29
*** moshele has quit IRC19:29
*** banix has joined #openstack-neutron19:30
*** fzdarsky|afk has quit IRC19:42
*** mihero has quit IRC19:42
*** mihero has joined #openstack-neutron19:43
*** tflynn has joined #openstack-neutron19:45
*** pcaruana has quit IRC19:52
*** ilyashakhat_mobi has joined #openstack-neutron19:57
*** yamahata has joined #openstack-neutron19:57
*** tflynn has quit IRC19:59
*** markvoelker has joined #openstack-neutron20:00
*** SKipp_ has quit IRC20:01
*** oomichi- has joined #openstack-neutron20:02
*** danofsatx-ghost has joined #openstack-neutron20:03
*** magic has joined #openstack-neutron20:04
*** magic is now known as Guest3621520:04
*** markvoelker has quit IRC20:05
*** Anticime1 has joined #openstack-neutron20:05
*** evrardjp_ has joined #openstack-neutron20:05
*** ishafran has joined #openstack-neutron20:06
*** thorst has joined #openstack-neutron20:07
*** banix has quit IRC20:08
openstackgerritMohit Malik proposed openstack/neutron: Integrate Address Scope OVO into DB  https://review.openstack.org/30800520:08
*** HenryG_ has joined #openstack-neutron20:08
*** briancurtin has quit IRC20:10
*** csatari has quit IRC20:10
*** xiaoyang has quit IRC20:10
*** StevenK has quit IRC20:10
*** oomichi has quit IRC20:10
*** xinliang has quit IRC20:10
*** evrardjp has quit IRC20:10
*** HenryG has quit IRC20:10
*** Anticimex has quit IRC20:10
*** nirmoy has quit IRC20:10
*** amit213 has quit IRC20:10
*** scottda has quit IRC20:10
*** mrvinti has quit IRC20:10
*** dlundquist has quit IRC20:10
*** vern has quit IRC20:10
*** kgalanov has quit IRC20:10
*** danofsatx has quit IRC20:10
*** evrardjp_ is now known as evrardjp20:10
*** danofsatx-ghost is now known as danofsatx20:10
*** slaweq__ has joined #openstack-neutron20:11
*** briancurtin has joined #openstack-neutron20:11
*** csatari has joined #openstack-neutron20:11
*** StevenK has joined #openstack-neutron20:11
*** nirmoy has joined #openstack-neutron20:11
*** amit213 has joined #openstack-neutron20:11
*** scottda has joined #openstack-neutron20:11
*** mrvinti has joined #openstack-neutron20:11
*** dlundquist has joined #openstack-neutron20:11
*** vern has joined #openstack-neutron20:11
*** kgalanov has joined #openstack-neutron20:11
*** briancurtin has quit IRC20:12
*** csatari has quit IRC20:12
*** xinliang has joined #openstack-neutron20:12
*** thorst has quit IRC20:13
*** saggi has quit IRC20:13
*** saggi has joined #openstack-neutron20:15
*** sdake has joined #openstack-neutron20:15
*** kgalanov has quit IRC20:16
*** saggi has quit IRC20:16
*** amit213 has quit IRC20:16
*** jlibosva has quit IRC20:18
*** briancurtin has joined #openstack-neutron20:18
*** slaweq__ has quit IRC20:19
openstackgerritJohn Schwarz proposed openstack/neutron: l3 ha: don't send routers without '_ha_interface'  https://review.openstack.org/36565320:19
*** slaweq has quit IRC20:20
*** jlibosva has joined #openstack-neutron20:21
*** slaweq has joined #openstack-neutron20:21
*** ilyashakhat_mobi has quit IRC20:24
*** sdake has quit IRC20:25
*** briancurtin has quit IRC20:25
*** kgalanov has joined #openstack-neutron20:25
*** amit213 has joined #openstack-neutron20:25
*** csatari has joined #openstack-neutron20:25
*** marst has quit IRC20:26
*** briancurtin has joined #openstack-neutron20:26
prometheanfireguess everyone is gone now20:26
*** salv-orlando has joined #openstack-neutron20:29
*** yfried has joined #openstack-neutron20:30
*** salv-orl_ has quit IRC20:32
*** nmagnezi has joined #openstack-neutron20:33
*** Jeffrey4l has quit IRC20:34
*** clenimar has quit IRC20:35
*** nyechiel_ has joined #openstack-neutron20:35
*** Jeffrey4l has joined #openstack-neutron20:36
*** dkehn has quit IRC20:36
*** dkehn_ has quit IRC20:36
*** john-davidge has quit IRC20:37
*** Alex_Stef has quit IRC20:52
*** nplanel has joined #openstack-neutron20:52
*** Anticime1 is now known as Anticimex20:53
*** banix has joined #openstack-neutron20:57
*** banix has quit IRC20:59
*** banix has joined #openstack-neutron21:00
openstackgerritJakub Libosvar proposed openstack/neutron: trunk: Handle subports notification in ovs agent side  https://review.openstack.org/36379821:12
*** jckasper has joined #openstack-neutron21:13
*** nplanel has quit IRC21:15
*** xdcc has quit IRC21:19
*** jckasper has quit IRC21:19
*** john-davidge has joined #openstack-neutron21:20
*** jckasper has joined #openstack-neutron21:20
*** fragatina has joined #openstack-neutron21:20
openstackgerritJakub Libosvar proposed openstack/neutron: trunk: Handle subports notification in ovs agent side  https://review.openstack.org/36379821:21
openstackgerritJakub Libosvar proposed openstack/neutron: OVSDBHandler for trunk ports  https://review.openstack.org/34746621:21
*** fragatina has quit IRC21:24
*** fragatina has joined #openstack-neutron21:24
*** jckasper has quit IRC21:25
*** fragatina has quit IRC21:25
*** fragatina has joined #openstack-neutron21:25
*** clenimar has joined #openstack-neutron21:28
*** buttercup has joined #openstack-neutron21:28
*** thorst has joined #openstack-neutron21:30
*** nmagnezi has quit IRC21:33
*** sdake has joined #openstack-neutron21:34
*** jckasper has joined #openstack-neutron21:39
*** thorst has quit IRC21:39
*** nyechiel_ has quit IRC21:41
*** jckasper has quit IRC21:41
*** jlibosva has quit IRC21:44
*** banix has quit IRC21:45
*** iranzo has quit IRC21:51
*** sdake has quit IRC21:52
*** nplanel has joined #openstack-neutron21:53
*** obondarev has joined #openstack-neutron21:59
*** markvoelker has joined #openstack-neutron22:00
*** akshai has joined #openstack-neutron22:02
*** akshai has quit IRC22:03
*** fragatina has quit IRC22:04
*** markvoelker has quit IRC22:06
*** john-davidge has quit IRC22:06
*** buttercup has quit IRC22:07
*** buttercup has joined #openstack-neutron22:13
*** jckasper has joined #openstack-neutron22:15
*** akshai has joined #openstack-neutron22:16
*** jckasper has quit IRC22:19
*** zhugaoxiao has quit IRC22:19
*** zhugaoxiao has joined #openstack-neutron22:19
*** obondarev has quit IRC22:20
*** marst has joined #openstack-neutron22:21
*** sindhude__ has joined #openstack-neutron22:34
*** thorst has joined #openstack-neutron22:36
*** sindhude__ has quit IRC22:41
*** claudiub|2 has quit IRC22:44
*** thorst has quit IRC22:45
*** vishwanathj has joined #openstack-neutron22:55
*** john-davidge has joined #openstack-neutron22:55
*** vishwanathj has quit IRC22:55
*** john-davidge has quit IRC23:00
*** adriant has joined #openstack-neutron23:02
*** eezhova has quit IRC23:07
*** yfried has quit IRC23:08
*** buttercup has quit IRC23:14
*** john-davidge has joined #openstack-neutron23:15
*** john-davidge has quit IRC23:17
*** akshai has quit IRC23:18
*** baoli has joined #openstack-neutron23:30
*** baoli has quit IRC23:30
*** ishafran has quit IRC23:30
*** john-davidge has joined #openstack-neutron23:33
*** john-davidge has quit IRC23:38
*** yamahata has quit IRC23:38
*** lujinluo has joined #openstack-neutron23:39
openstackgerritLujin Luo proposed openstack/neutron: Add a composite unique key to floatingip table in Neutron database  https://review.openstack.org/27601823:40
*** thorst has joined #openstack-neutron23:41
*** jlibosva has joined #openstack-neutron23:46
*** gongysh has joined #openstack-neutron23:46
*** thorst has quit IRC23:49
*** buttercup has joined #openstack-neutron23:49
*** jlibosva has quit IRC23:51
*** john-davidge has joined #openstack-neutron23:52
*** lujinluo has quit IRC23:56
*** john-davidge has quit IRC23:56
*** lujinluo has joined #openstack-neutron23:57

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