Thursday, 2017-09-07

*** fzdarsky__ has joined #openstack-neutron-ovn01:49
*** fzdarsky_ has quit IRC01:53
*** yamamoto_ has joined #openstack-neutron-ovn02:09
*** rpb has quit IRC03:34
*** anilvenkata has joined #openstack-neutron-ovn04:05
*** yamamot__ has joined #openstack-neutron-ovn04:20
*** yamamoto_ has quit IRC04:21
*** kevinbenton has quit IRC04:21
*** kevinbenton has joined #openstack-neutron-ovn04:23
*** janki has joined #openstack-neutron-ovn04:55
*** yamamot__ has quit IRC05:07
*** yamamoto_ has joined #openstack-neutron-ovn05:07
*** pcaruana has joined #openstack-neutron-ovn05:15
*** anilvenkata has quit IRC05:39
*** zefferno has joined #openstack-neutron-ovn06:10
*** ajo has quit IRC06:34
openstackgerritxurong00037997 proposed openstack/networking-ovn master: DHCP options for subnet synchronize each time when run ovn db sync or restart neutron-server, it is because there is no classless_static_route for metadata.
*** lucas-afk is now known as lucasagomes08:31
lucasagomesnumans, apparently the patch didn't work09:09
lucasagomesnumans, I think we will need to change tempest to accomodate that missing ping from ovn09:09
numanslucasagomes, :(. i am looking into the tempest changes.09:11
lucasagomescool, we can add a hacky patch just to see if that's the cause of the job being failing09:11
lucasagomeslike just fake the result of that method to always succeed09:11
lucasagomesif it works we can then upload a proper patch09:11
openstackgerritxurong00037997 proposed openstack/networking-ovn master: DHCP options for subnet synchronize each time
*** yamamoto_ has quit IRC09:28
*** ajo has joined #openstack-neutron-ovn09:45
*** yamamoto has joined #openstack-neutron-ovn09:46
*** anilvenkata has joined #openstack-neutron-ovn09:49
*** yamamoto has quit IRC10:14
openstackgerritMerged openstack/networking-ovn master: Set requested-chassis with binding host_id.
*** yamamoto has joined #openstack-neutron-ovn10:19
*** livelace has joined #openstack-neutron-ovn10:23
numanslucasagomes, fingeres crossed -
* lucasagomes looks10:50
lucasagomesnumans, good stuff on forking tempest for the test (-:10:51
numanslucasagomes, luckily ooo-extras provided the option :)10:51
numanslucasagomes, here is the tempest clone -
numanslucasagomes, i have by passed the ping validation and put some prints. hope it passes now :)10:53
lucasagomescool :D10:55
*** ajo has quit IRC10:57
numanslucasagomes, if you grep for 'run_validation' it is set to False10:59
numansi wonder if we should set it to true10:59
numansany thoughts ?10:59
lucasagomesnumans, I would have to dig in the tempest code to see exactly what that config does to be honest10:59
lucasagomesby it's name it sounds like yeah we want to run the validations11:00
lucasagomesbut, config names in openstack can be misleading sometimes heh11:00
* lucasagomes looks 11:00
lucasagomesnumans, it has always been false ?11:01
lucasagomescause the patch increasing the ping count uses validations, no ?11:01
numanslucasagomes, yes.11:02
numanslucasagomes, in the case of tripleo yes, its set to true11:02
numanslucasagomes, let me check11:02
numanslucasagomes, ignore my comment,11:03
numanslucasagomes, i think its ok :)11:03
numansgot confused with the names. still confused11:04
lucasagomeshah ok we will find out in a few anyway11:04
lucasagomesI hear you, sometimes even the help string does not help much11:04
numanslucasagomes, if you see this log in networking-ovn job, we see the same ERROR but still the tests pass11:06
lucasagomesotherwiseguy, around ? Am I doing something wrong here or the ovsdb locks does not work ?
lucasagomesnumans, ^ if you have a min11:07
lucasagomesnumans, looking11:07
numanslucasagomes, i think its a bit early for otherwiseguy :)11:07
lucasagomesnumans, hmmmm strange yeah... It's hard cause I don't see any other explicit errors on that oooq run11:08
lucasagomesso I thought the ping could have something to deal with it11:08
lucasagomesbut now I don't know, if this other tests is passing even when pinging is failing11:08
numanslucasagomes, i would suggest you enable the debug logs in ovsdb-server and monitor the log, you would clearly see a json rpc message to acquire the lock11:08
lucasagomesnumans, lemme do it11:09
numanslucasagomes, you could use ovs-appclt -t <CONTROL_SOCKET_PATH> vlog/set dbg11:09
numansi suppose11:09
numansyour test app exits after the print ?11:09
numanslucasagomes, may be you can put  a while loop or sleep of 5 seconds and then see the status11:10
lucasagomesI'm just checking whether it got the lock or not11:10
lucasagomeshmm good point11:10
lucasagomeslemme see11:10
numanslucasagomes, it would take some time for the lock reply11:10
lucasagomesnumans, oh right11:10
lucasagomesyeah that could be it then11:11
lucasagomesthanks numans !11:11
lucasagomesnumans, yup does work :D11:13
numanslucasagomes, cool11:13
lucasagomesnumans, hah I will dig into it but, something is odd about this locks:
lucasagomesbasically both idls got the lock with the same name11:19
lucasagomesthere's a race there  maybe ?!11:19
numanslucasagomes, i think the problem is both of them have the same IDL connection :)11:19
numanslucasagomes, its as per design :)11:20
numanslucasagomes, let me point you to the code11:20
* lucasagomes checks the get_connection()11:20
lucasagomescause I'm trying to figure out how I'm going to use the locks for the maintenance thread11:20
lucasagomesoh gotcha11:21
lucasagomesoh right it's a class attribute11:22
numanslucasagomes, you have to either fork or run two test apps11:22
*** amuller has joined #openstack-neutron-ovn11:22
* lucasagomes thinks how he will use these locks now11:22
*** lucasagomes is now known as lucas-hungry11:32
*** lucas-hungry is now known as lucasagomes12:10
*** ajo has joined #openstack-neutron-ovn12:26
*** jamesdenton has quit IRC12:27
*** jamesdenton has joined #openstack-neutron-ovn12:33
*** mmichelson has joined #openstack-neutron-ovn12:42
*** mmichelson has quit IRC12:43
*** mmichelson has joined #openstack-neutron-ovn12:46
lucasagomesnumans, I said I didn't need that patch to acquire multiple locks with different names in the same idl but, I'm hitting it now with my code heh I probably will work around it another way12:59
lucasagomesbut yeah, now ovsdb_monitor and the maintenance thread is trying to get the lock12:59
lucasagomesand end up overwritting one another12:59
numanslucasagomes, so do you need that feature ?13:00
numanslucasagomes, or you have a different way to address ?13:00
lucasagomesnumans, well it would come handy with my current design13:00
lucasagomesdunno if I actually need it13:00
lucasagomesI will see13:00
lucasagomesI was reusing the same idl I had on the mech driver13:00
lucasagomesI could probably just instantiate a new one, but, let's see...13:01
* lucasagomes wants to avoid create moar in-memory replicas13:01
russellbHi everyone ... this has come up a few times before, but we're long overdue for retiring this IRC channel and just using #openstack-neutron ... so let's do that starting now.13:07
russellbhm ... I don't have ops in here anymore13:08
russellbwill look into how to get ops to update topic ...13:08
*** ChanServ sets mode: +o russellb13:09
*** russellb changes topic to "THIS CHANNEL IS DEPRECATED. Please use #openstack-neutron for all OpenStack+OVN discussion."13:09
*** russellb sets mode: +m 13:14
*** nmagnezi has joined #openstack-neutron-ovn13:15
*** otherwiseguy has left #openstack-neutron-ovn13:27
*** amuller has left #openstack-neutron-ovn13:28
*** otherwiseguy has joined #openstack-neutron-ovn13:28
*** jamesdenton has left #openstack-neutron-ovn13:44
*** thegreenhundred has joined #openstack-neutron-ovn13:50
*** janki has quit IRC14:36
*** anilvenkata is now known as anilvenkata|AFK14:45
*** otherwiseguy has left #openstack-neutron-ovn15:04
*** irenab has quit IRC15:12
*** irenab has joined #openstack-neutron-ovn15:12
*** mmichelson has left #openstack-neutron-ovn15:44
*** anilvenkata|AFK has quit IRC16:13
*** zefferno has quit IRC16:34
*** lucasagomes has left #openstack-neutron-ovn16:39
*** ajo has quit IRC16:55
*** tongl has joined #openstack-neutron-ovn17:26
*** livelace has quit IRC17:38
*** livelace has joined #openstack-neutron-ovn17:50
*** livelace has quit IRC17:56
*** livelace has joined #openstack-neutron-ovn18:13
*** amuller has joined #openstack-neutron-ovn18:43
*** amuller has quit IRC19:14
*** amuller has joined #openstack-neutron-ovn19:17
*** amuller has quit IRC19:28
*** pcaruana has quit IRC19:31
*** dmellado has quit IRC19:45
*** ajo has joined #openstack-neutron-ovn20:22
*** ajo has quit IRC21:36
*** amuller has joined #openstack-neutron-ovn21:58
*** thegreenhundred has quit IRC22:11
*** amuller has left #openstack-neutron-ovn22:35
*** tongl has quit IRC23:25

Generated by 2.15.3 by Marius Gedminas - find it at!