Sunday, 2019-02-17

*** hongbin has quit IRC00:03
cgoncalvesI didn't notice that particularly. I did, though, notice that both our act-stby patches failed with exact same error00:28
cgoncalvesand yours passed a few days ago so something got it broken00:29
johnsomI will try to take a look at them both tomorrow.  It's tapas night here.00:44
cgoncalvesenjoy :)01:14
*** sapd1 has joined #openstack-lbaas01:24
johnsomI made potato bravas:02:52
johnsomhttps://usercontent.irccloud-cdn.com/file/5IwTS7Zf/1550371966.JPG02:52
johnsomThe wife made gambes02:54
johnsomhttps://usercontent.irccloud-cdn.com/file/EGv1DxNv/1550372073.JPG02:54
*** sapd1 has quit IRC02:56
*** gcheresh has joined #openstack-lbaas06:39
*** AlexStaf has joined #openstack-lbaas08:50
openstackgerritzhulingjie proposed openstack/neutron-lbaas master: Update hacking version to 1.1.0  https://review.openstack.org/62864609:04
openstackgerritzhulingjie proposed openstack/neutron-lbaas master: Update hacking version to 1.1.0  https://review.openstack.org/62864609:05
*** sapd1 has joined #openstack-lbaas09:46
cgoncalvesjohnsom, looks good! now I know where to go for a bite in Portland ;)09:50
*** AlexStaf has quit IRC11:33
*** oanson has quit IRC11:48
openstackgerritJacky Hu proposed openstack/octavia-dashboard master: Allow cascade deletion of load balancer  https://review.openstack.org/55338113:08
*** salmankhan has joined #openstack-lbaas13:08
*** psachin has joined #openstack-lbaas13:16
openstackgerritMerged openstack/octavia-dashboard master: Imported Translations from Zanata  https://review.openstack.org/63713813:17
*** psachin has quit IRC13:28
*** yboaron has joined #openstack-lbaas13:30
*** salmankhan has quit IRC13:47
openstackgerritJacky Hu proposed openstack/octavia master: Speed up pylint by using multiple cores  https://review.openstack.org/63734814:02
openstackgerritJacky Hu proposed openstack/octavia master: Speed up pylint by using multiple cores  https://review.openstack.org/63734814:06
openstackgerritJacky Hu proposed openstack/octavia master: Speed up pylint by using multiple cores  https://review.openstack.org/63734814:09
*** sapd1 has quit IRC14:18
*** sapd1 has joined #openstack-lbaas14:55
*** AlexStaf has joined #openstack-lbaas15:37
*** gcheresh has quit IRC15:39
*** sapd1 has quit IRC18:10
*** AlexStaf has quit IRC18:19
*** AlexStaf has joined #openstack-lbaas18:41
openstackgerritMerged openstack/octavia-dashboard master: Allow cascade deletion of load balancer  https://review.openstack.org/55338119:29
*** salmankhan has joined #openstack-lbaas21:34
nmagnezijohnsom, o/22:06
nmagnezijohnsom, here by any chance?22:06
johnsomnmagnezi Yes22:06
nmagnezijohnsom, hi :)22:06
nmagnezijohnsom, testing for  https://storyboard.openstack.org/#!/story/2004112#comment-11598222:06
nmagnezijohnsom, to double-check myself, I spawned a devstack node with ubuntu amp and executed those steps22:07
nmagnezijohnsom, for reference, since this shouldn't be an issue for ubuntu amps at the moment22:07
nmagnezijohnsom, but for some reason I see that an IPv6 address (and actually additional IPv4 addresses) are not added to the namespace NICs22:08
johnsomYeah, the current solution on the ubuntu side is to just bring up all of the subnets22:08
johnsomI think this test would fail if it wasn't working correctly: https://github.com/openstack/octavia-tempest-plugin/blob/master/octavia_tempest_plugin/tests/scenario/v2/test_ipv6_traffic_ops.py#L11322:10
nmagnezijohnsom, what an22:10
nmagnezijohnsom, sorry mistyped. Wanted to paste this http://paste.openstack.org/show/745232/22:10
nmagnezijohnsom, I'm. Not saying there's a bug, just checking what am I doing wrong at this point :)22:11
johnsomRight, I understand.22:11
johnsomCan you paste "sudo ip netns exec amphora-haproxy ip a" from inside the amp?22:11
nmagnezijohnsom, http://paste.openstack.org/show/745233/22:12
nmagnezijohnsom, ah. ip a.. on the way22:12
nmagnezijohnsom, http://paste.openstack.org/show/745234/22:12
johnsomNeat, ok how about a "openstack server list" and "openstack subnet list"22:14
nmagnezijohnsom, np. Just saying that those members are just random IPs I made up on those subnets.. there are no instances with those addresses22:15
nmagnezijohnsom, will that cause the amp agent not to configure the NIC?22:15
johnsomNo, it should still plug them.22:15
nmagnezijohnsom, the current eth config http://paste.openstack.org/show/745235/22:16
johnsomI'm looking for the subnets nova attached to the instance. That is why I am asking for the additional info22:16
nmagnezinp22:17
nmagnezihttp://paste.openstack.org/show/745236/22:17
nmagneziThe amp with 192.168.0.197 is the one I currently use22:17
johnsomCan you do a "openstack loadbalancer member show cce614ac-c685-4e2c-be32-4c4744e138c9 07da18ab-b567-46a3-aef8-de8d4ebfe6c322:21
nmagneziSure http://paste.openstack.org/show/745237/22:22
johnsomSo the VIP is on the same network?22:22
nmagneziThe VIP is 11.11.11.1922:23
johnsomWe may have another bug if it's the case where the VIP and the member are on the same network.22:23
johnsomYeah, ok, that is probably the case, we probably have another bug. I know some of the plug stuff is skipped if the network is already plugged for the VIP.22:24
johnsomI will open a story for that.22:24
nmagneziAye22:25
nmagneziI'm going to look more closely at this. IMO we need decouple the interface creation for Ubuntu an RH. They share some part that they don't (the baseOS stuff)22:26
nmagneziI will give it another spin where the members are on a different subnet from the VIP22:26
nmagneziBtw the networking for RHEL8 should eventually also look different (NetworkManager)22:27
nmagneziAt some point I'll add it too22:27
nmagneziBut first things first22:27
johnsomYeah, Ubuntu is changing to netplan, so we need to do some work there too22:27
openstackgerritMerged openstack/octavia-dashboard master: fix typo mistakes  https://review.openstack.org/62541722:29
johnsomReally I would like to re-work the whole member plug flow. It's not handling subnets the way it should IMO.22:29
johnsomhttps://storyboard.openstack.org/#!/story/200503822:30
nmagnezijohnsom, ^^ we should run some spell checker against are code bases and get this over with..22:30
johnsomI almost abandoned that given it was in a comment....22:30
nmagnezilol22:31
johnsomIt brings zero benefit to the community and just wastes gate resources22:31
nmagnezifully agree22:31
nmagneziThis really made me lol https://review.openstack.org/#/c/625417/1/octavia_dashboard/api/rest/lbaasv2.py@4422:31
nmagneziNever saw the mordred here before btw22:32
johnsomWe would have to train a aspell dictionary, but we could put a test in tox...22:32
nmagnezijohnsom, https://review.openstack.org/#/c/637074/1//COMMIT_MSG 20 minutes to boot a VM ?!22:40
nmagnezijohnsom, wow.. that's insane.. I run nested VMs for spawning test nodes (devstack etc) and it takes a couple of min22:40
johnsomYes, it has happened with nova on non-nested virtualization enabled nested compute hosts. It also happens on virtualbox22:41
johnsomYeah, it's only when qemu is running in tcg mode22:41
johnsomaccel=tcg22:41
nmagneziI can't recall if tripleO is overriding those defaults with some other values, but indeed its good to have those numbers down to 12022:43
*** eandersson has joined #openstack-lbaas22:43
*** salmankhan has quit IRC23:16
*** yboaron_ has joined #openstack-lbaas23:24
*** yboaron has quit IRC23:26

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!