| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add amphora statistics to the admin API https://review.openstack.org/585031 | 00:11 |
|---|---|---|
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Adds get_amphora_stats to the service client https://review.openstack.org/585060 | 00:31 |
| *** threestrands has quit IRC | 00:38 | |
| *** hongbin has joined #openstack-lbaas | 00:39 | |
| *** longkb has joined #openstack-lbaas | 00:40 | |
| korean101 | hmmmm | 00:44 |
| korean101 | johnsom: HI | 00:44 |
| korean101 | johnsom: I have a question | 00:44 |
| johnsom | korean101 Hi, I might have an answer | 00:45 |
| korean101 | johnsom: :) | 00:45 |
| korean101 | johnsom: lb-mgmt-network is a virtual network right? | 00:45 |
| johnsom | yes, a neutron network | 00:46 |
| korean101 | johnsom: how can octavia-worker connect to lb-mgmt-net (9443 port)? | 00:46 |
| korean101 | johnsom: Failed to establish a new connection: [Errno 113] No route to host | 00:46 |
| korean101 | johnsom: i have a some ERRORs | 00:47 |
| korean101 | johnsom: octavia-worker have a real network | 00:47 |
| korean101 | johnsom: real network <--> neutron network | 00:47 |
| johnsom | There are many ways to do it. Some people use provider networks, some add the instances to the vxlan, some add a port directly to OVS or the linux bridge locally (this is how devstack does it) | 00:47 |
| *** JudeC has quit IRC | 00:48 | |
| korean101 | johnsom: yes. i know devstack's way | 00:48 |
| johnsom | korean101 https://github.com/openstack/octavia/blob/master/devstack/plugin.sh#L341 | 00:48 |
| johnsom | Oh, ha, well, there is another link | 00:48 |
| korean101 | johnsom: but I use kolla-ansible | 00:49 |
| johnsom | OpenStack Ansible uses a provider VLAN | 00:49 |
| korean101 | johnsom: my octavia daemons in controller nodes | 00:49 |
| korean101 | johnsom: provider vlan? | 00:49 |
| korean101 | johnsom: can i get some links? | 00:49 |
| korean101 | johnsom: any documents? | 00:49 |
| johnsom | https://github.com/openstack/openstack-ansible-os_octavia/blob/master/tasks/octavia_mgmt_network.yml | 00:50 |
| johnsom | https://docs.openstack.org/neutron/pike/admin/intro-os-networking.html#provider-networks | 00:51 |
| johnsom | https://docs.openstack.org/openstack-ansible-os_octavia/latest/configure-octavia.html#setup-a-neutron-network-for-use-by-octavia | 00:52 |
| korean101 | johnsom: additional bridge needed? | 00:53 |
| johnsom | It totally depends on your environment and how you want to do your networking. | 00:54 |
| johnsom | Sadly, I cannot spend a lot of time tonight helping, I have some pressing work to finish. | 00:54 |
| korean101 | johnsom: ok | 00:54 |
| korean101 | johnsom: this links helps to me | 00:55 |
| korean101 | johnsom: really thanks!!! | 00:55 |
| korean101 | johnsom: always | 00:55 |
| johnsom | Oh good. | 00:55 |
| *** rcernin_ has joined #openstack-lbaas | 00:56 | |
| *** rcernin has quit IRC | 00:58 | |
| *** links has joined #openstack-lbaas | 01:17 | |
| *** yamamoto has joined #openstack-lbaas | 01:20 | |
| *** yamamoto has quit IRC | 01:24 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add amphora statistics to the admin API https://review.openstack.org/585031 | 01:39 |
| *** fnaval has joined #openstack-lbaas | 01:40 | |
| *** fnaval has quit IRC | 01:45 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Adds get_amphora_stats to the service client https://review.openstack.org/585060 | 01:46 |
| *** ramishra has joined #openstack-lbaas | 03:01 | |
| *** JudeC has joined #openstack-lbaas | 04:15 | |
| *** hongbin has quit IRC | 04:16 | |
| *** phuoc_ has joined #openstack-lbaas | 04:45 | |
| *** phuoc has quit IRC | 04:48 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Add an active/standby scenario test https://review.openstack.org/584681 | 05:03 |
| *** JudeC has quit IRC | 05:09 | |
| rm_work | cgoncalves: in that repo the latest version of git is git-2.18.0-1.WANdisco.402.x86_64.rpm from 03-Jul-2018 | 05:17 |
| rm_work | it isn't sorted by date :P | 05:17 |
| rm_work | but yeah i didn't even realize that they HAD a cent7 repo, but i think the version is essentially the same | 05:17 |
| rm_work | yeah | 05:18 |
| rm_work | it is | 05:18 |
| rm_work | but prolly i should anyway, maybe compiled against different libs | 05:18 |
| * rm_work shrugs | 05:18 | |
| johnsom | You don't have to etch your patch on a stone tablet to do a "git commit" anymore? | 05:18 |
| rm_work | not with this fancy new git tech from the turn of the century! :P | 05:19 |
| *** yboaron has joined #openstack-lbaas | 05:19 | |
| johnsom | I have a working VRRP scenario gate... | 05:19 |
| johnsom | At least it works local... | 05:19 |
| rm_work | ooo | 05:20 |
| rm_work | yeah i have one working internally | 05:20 |
| rm_work | err, i did, but i might need to rebase it a ton | 05:20 |
| johnsom | HA, and forgot to run pep on it... sigh | 05:20 |
| rm_work | but mine tracks the master, not sure how yours works | 05:20 |
| rm_work | (my system updates master/standby in the amp table) | 05:21 |
| johnsom | I added amphora stats API so I can see which one handled the traffic | 05:21 |
| rm_work | ahhh nice | 05:21 |
| rm_work | so that's how you do it | 05:21 |
| rm_work | clever | 05:21 |
| johnsom | Yeah, I am on a bit of a timeline, so... Shortest path from A->B | 05:22 |
| johnsom | Which happened to be extend the API | 05:22 |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Add an active/standby scenario test https://review.openstack.org/584681 | 05:22 |
| *** kobis has joined #openstack-lbaas | 05:25 | |
| *** kobis has quit IRC | 05:26 | |
| rm_work | lol | 05:26 |
| rm_work | k | 05:26 |
| *** kobis has joined #openstack-lbaas | 05:26 | |
| *** kobis has quit IRC | 05:33 | |
| johnsom | Calling it a night, catch you all tomorrow. | 05:45 |
| *** yamamoto has joined #openstack-lbaas | 05:46 | |
| *** yamamoto has quit IRC | 05:48 | |
| *** nmanos has joined #openstack-lbaas | 05:52 | |
| *** yamamoto has joined #openstack-lbaas | 05:56 | |
| korean101 | johnsom: see you~ | 05:59 |
| *** yamamoto has quit IRC | 06:00 | |
| *** kobis has joined #openstack-lbaas | 06:06 | |
| *** kobis has quit IRC | 06:06 | |
| *** kobis has joined #openstack-lbaas | 06:10 | |
| *** nmanos has quit IRC | 06:17 | |
| *** nmanos has joined #openstack-lbaas | 06:17 | |
| korean101 | hmmmmm | 06:19 |
| korean101 | octavia can't use DVR? | 06:19 |
| *** AlexStaf has joined #openstack-lbaas | 06:25 | |
| *** nmanos has quit IRC | 06:30 | |
| *** nmanos has joined #openstack-lbaas | 06:31 | |
| *** pcaruana has joined #openstack-lbaas | 06:34 | |
| *** ispp has joined #openstack-lbaas | 07:05 | |
| *** JudeC has joined #openstack-lbaas | 07:12 | |
| *** tesseract has joined #openstack-lbaas | 07:18 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 07:50 | |
| *** velizarx has joined #openstack-lbaas | 07:54 | |
| *** yboaron has quit IRC | 07:55 | |
| *** rcernin_ has quit IRC | 07:56 | |
| *** ftersin has quit IRC | 08:00 | |
| *** yamamoto has joined #openstack-lbaas | 08:09 | |
| *** rcernin_ has joined #openstack-lbaas | 08:09 | |
| *** JudeC has quit IRC | 08:19 | |
| *** kobis has quit IRC | 08:20 | |
| *** yamamoto has quit IRC | 08:26 | |
| *** phuoc_ has quit IRC | 08:27 | |
| *** kobis has joined #openstack-lbaas | 08:27 | |
| *** phuoc_ has joined #openstack-lbaas | 08:27 | |
| *** velizarx has quit IRC | 08:29 | |
| *** ispp has quit IRC | 08:30 | |
| *** velizarx has joined #openstack-lbaas | 08:31 | |
| *** ispp has joined #openstack-lbaas | 08:38 | |
| *** yboaron has joined #openstack-lbaas | 08:48 | |
| *** phuoc_ has quit IRC | 08:58 | |
| *** phuoc_ has joined #openstack-lbaas | 08:58 | |
| *** dmellado has quit IRC | 09:03 | |
| *** phuoc_ has quit IRC | 09:04 | |
| *** phuoc has joined #openstack-lbaas | 09:05 | |
| *** kiennt26 has joined #openstack-lbaas | 09:06 | |
| *** apuimedo has joined #openstack-lbaas | 09:11 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: DNM: Gate test with publish-openstack-docs-pti https://review.openstack.org/585185 | 09:25 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: DNM: Gate test with publish-openstack-docs-pti https://review.openstack.org/585185 | 09:38 |
| openstackgerrit | Adit Sarfaty proposed openstack/octavia master: Init provider drivers at app setup https://review.openstack.org/585193 | 09:39 |
| *** oanson has quit IRC | 09:45 | |
| *** oanson has joined #openstack-lbaas | 09:45 | |
| *** kiennt26 has quit IRC | 10:17 | |
| *** yboaron_ has joined #openstack-lbaas | 10:23 | |
| *** dmellado has joined #openstack-lbaas | 10:24 | |
| *** yboaron has quit IRC | 10:26 | |
| *** velizarx has quit IRC | 10:51 | |
| *** rcernin_ has quit IRC | 10:52 | |
| *** velizarx has joined #openstack-lbaas | 10:56 | |
| *** phuoc has quit IRC | 10:59 | |
| *** phuoc has joined #openstack-lbaas | 11:00 | |
| *** yamamoto_ has joined #openstack-lbaas | 11:14 | |
| *** yboaron_ has quit IRC | 11:17 | |
| *** yamamoto has joined #openstack-lbaas | 11:25 | |
| *** yamamoto_ has quit IRC | 11:27 | |
| *** atoth has joined #openstack-lbaas | 11:29 | |
| *** yamamoto has quit IRC | 11:29 | |
| *** yamamoto has joined #openstack-lbaas | 11:31 | |
| *** yamamoto has quit IRC | 11:32 | |
| *** yamamoto has joined #openstack-lbaas | 11:32 | |
| *** longkb has quit IRC | 11:39 | |
| *** kobis has quit IRC | 11:43 | |
| *** yamamoto has quit IRC | 11:47 | |
| *** phuoc has quit IRC | 11:51 | |
| *** phuoc has joined #openstack-lbaas | 11:52 | |
| *** yamamoto has joined #openstack-lbaas | 11:57 | |
| *** amuller has joined #openstack-lbaas | 11:58 | |
| *** yamamoto has quit IRC | 11:59 | |
| *** yamamoto has joined #openstack-lbaas | 12:01 | |
| *** strigazi has quit IRC | 12:13 | |
| *** strigazi has joined #openstack-lbaas | 12:14 | |
| *** yboaron has joined #openstack-lbaas | 12:26 | |
| *** yamamoto has quit IRC | 12:27 | |
| *** yamamoto has joined #openstack-lbaas | 12:30 | |
| crazik | goh. | 12:32 |
| crazik | What can be a reason for not using spare amphoras when new LB is being created? | 12:32 |
| crazik | each LB create request starts with amphora create, althought there already are spares... | 12:34 |
| *** ispp has quit IRC | 12:39 | |
| *** velizarx has quit IRC | 12:53 | |
| *** velizarx has joined #openstack-lbaas | 12:58 | |
| *** ispp has joined #openstack-lbaas | 12:59 | |
| cgoncalves | crazik, spare amps are used when a failover occurs | 13:04 |
| johnsom | korean101: we do work on networks using DVR however DVR has had a lot of bugs that have caused some things to not work. For example floating IPs were broken with DVR until Pike | 13:06 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Fixing Octavia generating json named file containing yaml https://review.openstack.org/585024 | 13:06 |
| *** ianychoi has quit IRC | 13:08 | |
| johnsom | crazik: If you have anti-affinity enabled it will not use the spares pool as nova cannot guarantee the spare will meet the anti-affinity rules | 13:08 |
| *** phuoc has quit IRC | 13:08 | |
| *** ianychoi has joined #openstack-lbaas | 13:08 | |
| *** phuoc has joined #openstack-lbaas | 13:08 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Add an active/standby scenario test https://review.openstack.org/584681 | 13:26 |
| johnsom | Ha, I guess it helps if I tell the gate jobs to run the new tests and not the old ones...... | 13:26 |
| *** apuimedo is now known as apuimedo|afk | 13:52 | |
| *** fnaval has joined #openstack-lbaas | 13:55 | |
| *** KeithMnemonic has joined #openstack-lbaas | 14:24 | |
| *** yamamoto has quit IRC | 14:25 | |
| *** yamamoto has joined #openstack-lbaas | 14:27 | |
| *** yamamoto has quit IRC | 14:27 | |
| *** yamamoto has joined #openstack-lbaas | 14:27 | |
| *** links has quit IRC | 14:31 | |
| *** yamamoto has quit IRC | 14:43 | |
| *** yamamoto has joined #openstack-lbaas | 14:54 | |
| *** yboaron_ has joined #openstack-lbaas | 14:57 | |
| *** yboaron has quit IRC | 15:00 | |
| *** rcernin_ has joined #openstack-lbaas | 15:04 | |
| *** ispp has quit IRC | 15:26 | |
| *** rcernin_ has quit IRC | 15:30 | |
| *** pcaruana has quit IRC | 15:32 | |
| *** ispp has joined #openstack-lbaas | 15:41 | |
| *** hongbin has joined #openstack-lbaas | 15:42 | |
| hongbin | folks, if you find your gate break due to the recent landed validation logic in neutron, you might consider merging this fix: https://review.openstack.org/#/c/584000/ | 15:43 |
| *** velizarx has quit IRC | 15:53 | |
| *** ispp has quit IRC | 16:04 | |
| *** Krast has joined #openstack-lbaas | 16:18 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Add an active/standby scenario test https://review.openstack.org/584681 | 16:25 |
| johnsom | Oye, and remember to add the depends-on | 16:26 |
| johnsom | hongbin Thanks! | 16:28 |
| *** yamamoto has quit IRC | 16:33 | |
| eandersson | What is the heartbeat_key ? | 16:39 |
| eandersson | Or rather... setting it to insecure seems like a bad idea right? | 16:40 |
| eandersson | e.g. https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/octavia/templates/octavia.conf.j2#L56 | 16:40 |
| johnsom | Yes, that is just a "default" password worded to be a red-flag for folks. | 16:41 |
| johnsom | The heartbeat message that comes from the amps is HMAC encoded (also has a nonce code of the amp ID) with this key. | 16:41 |
| johnsom | So, summary, change it at deploy time. Changing it with running amps can lead to failovers | 16:42 |
| *** sanfern has joined #openstack-lbaas | 16:43 | |
| eandersson | I figured as much, but we tend to use kolla-ansible as a template, so wanted to double check :p | 16:45 |
| eandersson | Thanks johnsom | 16:45 |
| johnsom | NP | 16:45 |
| *** JudeC has joined #openstack-lbaas | 16:51 | |
| *** AlexeyAbashkin has quit IRC | 17:11 | |
| rm_work | nice, we're also using kolla-ansible for deployments soon ;) | 17:16 |
| *** ramishra has quit IRC | 17:17 | |
| *** tesseract has quit IRC | 17:17 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Fix DIB_REPOREF_amphora_agent not set on Git !=1.8.5 https://review.openstack.org/584856 | 17:26 |
| rm_work | cgoncalves: why is it "!=1.8.5"? | 17:34 |
| rm_work | was there literally a specific feature ONLY on 1.8.5? | 17:34 |
| rm_work | or should it just be >=1.8.5? | 17:34 |
| *** sanfern has quit IRC | 17:35 | |
| *** abaindur has joined #openstack-lbaas | 17:59 | |
| *** amuller has quit IRC | 18:41 | |
| crazik | cgoncalves: I was quite sure it uses spare amphoras not only for failover.. need to test on dev env again :/ | 18:55 |
| crazik | johnsom: hm, it's not the issue in this case | 18:55 |
| cgoncalves | rm_work, git -C only works >=1.8.5 | 18:56 |
| crazik | johnsom: I have anti-affinity enabled | 18:56 |
| rm_work | right, ok | 18:56 |
| rm_work | you have "!=" tho | 18:56 |
| rm_work | so was really confused what was specific about that one version :P | 18:56 |
| rm_work | probably you mean "Fix DIB_REPOREF_amphora_agent not set on Git <1.8.5" | 18:57 |
| *** yamamoto has joined #openstack-lbaas | 19:14 | |
| *** yamamoto has quit IRC | 19:19 | |
| johnsom | crazik: yes, so with anti-affinity enabled, spares pool is disabled for allocation. | 19:41 |
| *** yboaron_ has quit IRC | 19:47 | |
| cgoncalves | rm_work, nop. with the code as-is, on >1.8.5 the command would be 'git -c' (lowercase c) which doesn't work | 19:47 |
| cgoncalves | rm_work, 'git -C' (uppercase c) is only used when ==1.8.5 | 19:48 |
| cgoncalves | rm_work, check the 'Fedora 28' part on https://storyboard.openstack.org/#!/story/2003067 | 19:49 |
| *** rtjure has quit IRC | 19:57 | |
| colby_ | Hey All. Im setting up octavia using a ml2 gre network. I created the ports in neutron and create the bridge port on the node with the following command on our network nodes: | 20:16 |
| colby_ | ovs-vsctl -- --may-exist add-port br-int o-lbmgmt0 -- set Interface o-lbmgmt0 type=internal -- set Interface o-lbmgmt0 external-ids:iface-status=active -- set Interface o-lbmgmt0 external-ids:attached-mac=fa:16:3e:f4:c9:97 -- set Interface o-lbmgmt0 external-ids:iface-id=db98bbe6-ae9c-4cb1-ab2e-72c0e3cde841 -- set Interface o-lbmgmt0 external-ids:skip_cleanup=true | 20:16 |
| colby_ | how do I make that persistent on boot on Centos? | 20:16 |
| cgoncalves | colby_, create /etc/sysconfig/network-scripts/ifcfg-o-lbmgmt0 | 20:18 |
| *** rcernin_ has joined #openstack-lbaas | 20:18 | |
| cgoncalves | a template: https://github.com/openstack/tripleo-common/blob/master/playbooks/roles/octavia-controller-config/templates/ifcfg.j2 | 20:18 |
| johnsom | Wow, learn something new. I thought you had to shove it in ifup-local. Cool! | 20:19 |
| colby_ | Great! Thats exactly what I needed. Im not familiar with ovs network scripts. So was not sure on proper syntax | 20:19 |
| crazik | johnsom: hm, can I change this behavior? I thought spare pool is for faster LB deployment. In this case it would be only used for failover when compute node fails... | 20:30 |
| johnsom | crazik Spares pool is used for both new LBs and failover. however, if you enable anti-affinity we cannot pull an instance from the spares pool as nova can't guarantee the instances from the spares pool will meet the anti-affinity rules. This is probably an area we could enhance to work around the nova limitations | 20:32 |
| crazik | johnsom: but if I have 3 AZ, and 3 spares in pool.. | 20:33 |
| crazik | then spares could use anti-affinity and be ready for a new LB | 20:33 |
| *** nmanos has quit IRC | 20:35 | |
| crazik | in fact, 2 spares in pool should be enough for active/standby layout | 20:35 |
| *** crazik has quit IRC | 20:45 | |
| *** rtjure has joined #openstack-lbaas | 20:58 | |
| *** rcernin_ has quit IRC | 21:30 | |
| *** JudeC_ has joined #openstack-lbaas | 21:44 | |
| *** JudeC has quit IRC | 21:44 | |
| johnsom | crazik It is enough, but we can't add them to an anti-affinity server group in nova. It doesn't live migrate if the server being added doesn't meet the anti-affinity rules. We don't know "where" the spares compute instances are in the cloud as anti-affinity can mean different things in different clouds. By default it is just two different compute hosts, but it could be AZs, etc. | 21:44 |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Add an active/standby scenario test https://review.openstack.org/584681 | 21:48 |
| *** yamamoto has joined #openstack-lbaas | 21:59 | |
| *** yamamoto has quit IRC | 22:04 | |
| *** yamamoto has joined #openstack-lbaas | 22:06 | |
| *** fnaval has quit IRC | 22:21 | |
| *** rcernin has joined #openstack-lbaas | 22:30 | |
| *** hongbin has quit IRC | 22:30 | |
| *** fnaval has joined #openstack-lbaas | 23:00 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!