| *** rcernin has quit IRC | 00:07 | |
| *** sapd1_x has quit IRC | 00:41 | |
| *** rcernin has joined #openstack-lbaas | 01:05 | |
| *** altlogbot_3 has quit IRC | 02:12 | |
| *** altlogbot_3 has joined #openstack-lbaas | 02:13 | |
| openstackgerrit | Merged openstack/octavia stable/stein: Performance improvement for non-udp health checks https://review.opendev.org/659210 | 03:13 |
|---|---|---|
| *** goldyfruit has quit IRC | 03:25 | |
| *** psachin has joined #openstack-lbaas | 03:31 | |
| *** ricolin has joined #openstack-lbaas | 03:31 | |
| *** gcheresh_ has joined #openstack-lbaas | 04:32 | |
| *** gcheresh_ has quit IRC | 04:51 | |
| *** pcaruana has joined #openstack-lbaas | 05:08 | |
| *** blake has joined #openstack-lbaas | 05:13 | |
| *** pcaruana has quit IRC | 05:16 | |
| *** pcaruana has joined #openstack-lbaas | 05:17 | |
| *** gcheresh_ has joined #openstack-lbaas | 05:20 | |
| *** sapd1_x has joined #openstack-lbaas | 05:23 | |
| *** gcheresh_ has quit IRC | 05:27 | |
| *** ramishra has joined #openstack-lbaas | 05:32 | |
| *** sapd1_x has quit IRC | 05:35 | |
| *** blake has quit IRC | 05:37 | |
| *** sapd1_x has joined #openstack-lbaas | 06:05 | |
| *** sapd1_x has quit IRC | 06:18 | |
| openstackgerrit | inspurericzhang proposed openstack/octavia-tempest-plugin master: Replace git.openstack.org URLs with opendev.org URLs https://review.opendev.org/661191 | 06:24 |
| *** gthiemonge has quit IRC | 06:27 | |
| *** gthiemonge has joined #openstack-lbaas | 06:29 | |
| *** ramishra_ has joined #openstack-lbaas | 06:37 | |
| *** ramishra has quit IRC | 06:37 | |
| *** ltomasbo has joined #openstack-lbaas | 06:42 | |
| *** ramishra_ has quit IRC | 06:45 | |
| *** ramishra has joined #openstack-lbaas | 06:47 | |
| *** ivve has joined #openstack-lbaas | 06:52 | |
| *** ccamposr__ has quit IRC | 07:09 | |
| *** ccamposr has joined #openstack-lbaas | 07:10 | |
| *** icey has quit IRC | 07:34 | |
| *** mnaser has quit IRC | 07:35 | |
| *** fyx has quit IRC | 07:36 | |
| *** rcernin has quit IRC | 07:36 | |
| *** beisner has quit IRC | 07:39 | |
| *** fyx has joined #openstack-lbaas | 07:43 | |
| *** mnaser has joined #openstack-lbaas | 07:43 | |
| *** beisner has joined #openstack-lbaas | 07:43 | |
| *** mnaser has quit IRC | 08:02 | |
| *** mnaser has joined #openstack-lbaas | 08:05 | |
| *** rpittau|afk is now known as rpittau | 08:08 | |
| *** rm_work has quit IRC | 08:09 | |
| *** rm_work has joined #openstack-lbaas | 08:10 | |
| *** fyx has quit IRC | 08:10 | |
| *** fyx has joined #openstack-lbaas | 08:13 | |
| *** icey has joined #openstack-lbaas | 08:17 | |
| *** trident has quit IRC | 08:33 | |
| *** trident has joined #openstack-lbaas | 08:34 | |
| *** tesseract has joined #openstack-lbaas | 08:48 | |
| *** gthiemonge has quit IRC | 09:12 | |
| *** gthiemonge has joined #openstack-lbaas | 09:13 | |
| *** sapd1_x has joined #openstack-lbaas | 09:14 | |
| *** ut2k3 has joined #openstack-lbaas | 09:16 | |
| *** gcheresh_ has joined #openstack-lbaas | 09:19 | |
| ut2k3 | Hi johnsom :) just to give you feedback => that was the problem that the ports got mixed up. | 09:20 |
| *** gcheresh_ has quit IRC | 09:36 | |
| openstackgerrit | Merged openstack/octavia stable/rocky: Performance improvement for non-udp health checks https://review.opendev.org/659211 | 09:38 |
| *** sapd1_x has quit IRC | 09:46 | |
| *** tesseract has quit IRC | 09:56 | |
| *** tesseract has joined #openstack-lbaas | 09:59 | |
| *** tesseract has quit IRC | 10:00 | |
| *** tesseract has joined #openstack-lbaas | 10:01 | |
| *** ut2k3 has quit IRC | 10:21 | |
| *** luksky has quit IRC | 10:48 | |
| *** ramishra has quit IRC | 10:49 | |
| *** tesseract has quit IRC | 11:10 | |
| *** luksky has joined #openstack-lbaas | 11:17 | |
| *** psachin has quit IRC | 11:29 | |
| *** psachin has joined #openstack-lbaas | 11:38 | |
| *** jlaffaye_ has quit IRC | 11:49 | |
| *** jlaffaye has joined #openstack-lbaas | 11:49 | |
| *** ramishra has joined #openstack-lbaas | 11:58 | |
| *** psachin has quit IRC | 12:08 | |
| *** AlexStaf has quit IRC | 12:12 | |
| *** dims has left #openstack-lbaas | 12:19 | |
| *** ccamposr has quit IRC | 12:25 | |
| *** ccamposr has joined #openstack-lbaas | 12:26 | |
| *** ccamposr__ has joined #openstack-lbaas | 12:34 | |
| *** ccamposr has quit IRC | 12:37 | |
| openstackgerrit | Merged openstack/octavia stable/rocky: Treat null admin_state_up as False https://review.opendev.org/659257 | 12:53 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: [WIP] Transition member flows to use dicts https://review.opendev.org/657842 | 13:01 |
| *** goldyfruit has joined #openstack-lbaas | 13:06 | |
| *** luksky has quit IRC | 13:31 | |
| *** boden has joined #openstack-lbaas | 13:32 | |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia master: Add support for monitor_{address,port} in UDP members https://review.opendev.org/661308 | 14:43 |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia master: Fixed pool and members status with UDP loadbalancers https://review.opendev.org/661309 | 14:43 |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia master: WIP Fixed down server issue after reloading keepalived https://review.opendev.org/661310 | 14:43 |
| *** ivve has quit IRC | 14:45 | |
| *** boden has quit IRC | 14:47 | |
| *** gcheresh_ has joined #openstack-lbaas | 14:52 | |
| *** ccamposr has joined #openstack-lbaas | 14:59 | |
| *** ccamposr__ has quit IRC | 15:01 | |
| *** ccamposr__ has joined #openstack-lbaas | 15:02 | |
| *** ccamposr has quit IRC | 15:04 | |
| *** boden has joined #openstack-lbaas | 15:17 | |
| *** luksky has joined #openstack-lbaas | 15:21 | |
| *** rpittau is now known as rpittau|afk | 15:45 | |
| *** ivve has joined #openstack-lbaas | 15:55 | |
| *** gcheresh_ has quit IRC | 16:21 | |
| *** gcheresh_ has joined #openstack-lbaas | 16:31 | |
| *** ccamposr has joined #openstack-lbaas | 16:34 | |
| *** ccamposr__ has quit IRC | 16:36 | |
| *** ut2k3 has joined #openstack-lbaas | 16:47 | |
| ut2k3 | johnsom thanks again for your help yesterday :) | 16:48 |
| johnsom | ut2k3 You are welcome, glad you are back in business | 16:48 |
| *** ccamposr has quit IRC | 16:49 | |
| *** ut2k3 has quit IRC | 16:55 | |
| *** ricolin has quit IRC | 17:03 | |
| colin- | does nova support the concept of a "protected" or do-not-delete vm? | 17:20 |
| colin- | i promise that'll be the last nova Q i ask in here :) | 17:20 |
| ivve | colin-: openstack server lock <instance_id> | 17:22 |
| colin- | that's exactly it, thanks ivve | 17:23 |
| ivve | works with names too | 17:23 |
| colin- | "lock": {"locked_reason": "I don't want to work"} | 17:23 |
| colin- | oof | 17:23 |
| ivve | however it doesn't protect you from other admins :) | 17:24 |
| ivve | but its a good safeguard | 17:24 |
| *** gcheresh_ has quit IRC | 17:38 | |
| johnsom | How in the world can this ever be true and pass? https://github.com/openstack/octavia-tempest-plugin/blob/master/octavia_tempest_plugin/tests/scenario/v2/test_traffic_ops.py#L394 | 17:50 |
| *** pcaruana has quit IRC | 17:53 | |
| johnsom | Ah, I think I get it. Removing the HM would flip it to healthy in theory, since it's the alt port that is down. hnmm, this run the second member didn't go back into the pool, at least not in time for this test. | 17:54 |
| *** pcaruana has joined #openstack-lbaas | 18:48 | |
| *** pcaruana has quit IRC | 18:48 | |
| *** ccamposr has joined #openstack-lbaas | 18:51 | |
| *** luksky has quit IRC | 19:01 | |
| *** boden has quit IRC | 20:19 | |
| *** boden has joined #openstack-lbaas | 20:20 | |
| *** boden has quit IRC | 20:23 | |
| rm_work | johnsom: a taskflow `revert`'s "result" is *the return* from the `execute`, right? | 21:10 |
| rm_work | so if the execute() method returns "abc", the value of `result` that gets passed to revert() is "abc"? | 21:10 |
| johnsom | ummm, I *think* so | 21:10 |
| rm_work | k | 21:11 |
| johnsom | It is passed into the revert method. Pretty sure it's "result" | 21:11 |
| johnsom | https://docs.openstack.org/taskflow/latest/user/atoms.html#task | 21:12 |
| rm_work | grrr, vips in my devstack aren't working | 21:33 |
| rm_work | oh do i need a FLIP | 21:34 |
| rm_work | fff | 21:34 |
| johnsom | Ok, so somehow I broke health monitor delete. argh. It "compeletes" but doesn't remove it in the haproxy.cfg. | 21:37 |
| johnsom | Good news is tempest caught it. | 21:37 |
| rm_work | that's... actually fantastic newds | 21:38 |
| rm_work | *news | 21:38 |
| * johnsom high fives | 21:38 | |
| rm_work | ah i forgot to add my route | 21:45 |
| rm_work | k cool | 21:45 |
| rm_work | testing commencing :D | 21:45 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Allow multiple VIPs per LB https://review.opendev.org/660239 | 21:54 |
| johnsom | Ok, I see what is going on. It's some strange stuff we did in the flows due to the data models. Still not 100% sure why some of the other objects aren't failing to delete, but I know how to fix this in likely a cleaner way than we are doing it now. | 21:55 |
| johnsom | Yep, ok, this will actually clean up some tech debt. I like it. | 22:07 |
| *** ccamposr has quit IRC | 22:22 | |
| *** goldyfruit has quit IRC | 22:32 | |
| *** v1k0d3n has joined #openstack-lbaas | 22:54 | |
| v1k0d3n | can someone point me in the right direction? i tagged a built image with `--tag amphora`, but getting a really strange error in the octavia log about `Failed to build compute instance due to: Invalid input for field/attribute flavorRef. Value: . u'' is too short (HTTP 400)` | 22:55 |
| v1k0d3n | anyone know what i've done to mess this up? | 22:56 |
| *** goldyfruit has joined #openstack-lbaas | 22:56 | |
| johnsom | v1k0d3n That is an error from nova. It is trying to say that the nova flavor specified in the octavia.conf file is not valid. See this setting in your octavia.conf: https://docs.openstack.org/octavia/latest/configuration/configref.html#controller_worker.amp_flavor_id | 22:57 |
| johnsom | It needs to be set to a nova flavor ID (that the octavia account can see) that will be used for the amphora instances. | 22:58 |
| johnsom | Our devstack setup script does this: https://github.com/openstack/octavia/blob/master/devstack/plugin.sh#L433 | 22:58 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Allow multiple VIPs per LB https://review.opendev.org/660239 | 22:59 |
| johnsom | Though you might want a different size. rhel/centos images need at least 3GB disk | 22:59 |
| v1k0d3n | ah, so it needs to be `m1.amphora`? | 22:59 |
| johnsom | v1k0d3n You can name it whatever you want as long as the value in the config file matches what is in nova flavors list | 22:59 |
| johnsom | We just happened to call it m1.ampohra in our devstack scripts. It could be a uuid too | 23:00 |
| v1k0d3n | i don't know how i missed that! thank you so much. i thought i covered most of the docs fairly well too. | 23:00 |
| johnsom | We just pass the string through to nova, we don't even look at it | 23:01 |
| johnsom | Well, admittedly we need to work on installation docs..... | 23:01 |
| v1k0d3n | that makes sense. i can see why you would do that. | 23:01 |
| v1k0d3n | that's like most of us unfortunately. | 23:01 |
| v1k0d3n | i was trying to do this with kolla, and trying to find docs around project specific things in a "kolla way" can be very challenging. | 23:02 |
| v1k0d3n | sorry for one more question, but what is a decent size for small workloads typically? | 23:04 |
| v1k0d3n | this is just for a demo environment, really. for kubernetes workloads. | 23:04 |
| *** luksky has joined #openstack-lbaas | 23:04 | |
| *** rcernin has joined #openstack-lbaas | 23:04 | |
| johnsom | Copy what we have in the devstack script. If you build an ubuntu image, 2GB disk is fine. otherwise do 3GB. | 23:05 |
| johnsom | That will go up to a few thousand concurrent connections | 23:06 |
| v1k0d3n | that's great, and very useful info. thanks johnsom! | 23:06 |
| johnsom | No problem | 23:06 |
| *** rcernin has quit IRC | 23:11 | |
| *** goldyfruit has quit IRC | 23:20 | |
| v1k0d3n | johnsom: still seem to be encountering some other issue. that definitely helped though! | 23:25 |
| v1k0d3n | do you ever get an error about multiple networks, even when you're telling the os cli the network ID? | 23:26 |
| johnsom | No, can't say I have | 23:26 |
| johnsom | can you paste the output? | 23:27 |
| v1k0d3n | `Failed to build compute instance due to: Multiple possible networks found, use a Network ID to be more specific.` | 23:29 |
| johnsom | That would be this configuration setting in the octavia.conf: https://docs.openstack.org/octavia/latest/configuration/configref.html#controller_worker.amp_boot_network_list | 23:30 |
| v1k0d3n | at first i was using the subnet id, but i tried the network id just in case...seemed that i got the same error though. | 23:30 |
| johnsom | Almost always that should be just one network ID | 23:30 |
| v1k0d3n | that's the `amp_boot_network_list` just to be clear, right? | 23:31 |
| johnsom | This is mine: amp_boot_network_list = f2090d5b-ed97-488e-94ec-8ae852b4ed26 | 23:31 |
| v1k0d3n | the stein header on the docs page is pushing the hook down a little. | 23:31 |
| v1k0d3n | ah, ok. | 23:31 |
| v1k0d3n | i can try that. | 23:31 |
| v1k0d3n | now is that a flat network that you have there, or is that on the external/public side? | 23:32 |
| johnsom | It is the lb-mgmt-net, it can be any neutron network. The only catch is the controller processes need to be able to access that network. | 23:33 |
| johnsom | In many deployments it is a private neutron network, but some have put it on existing admin networks they already have or some use a "shared" service network for it. | 23:34 |
| johnsom | I don't know how kolla sets this network up. | 23:34 |
| v1k0d3n | lol setup...what is that you're talking about :) | 23:34 |
| v1k0d3n | i'm kidding...it might, but not without some changes that i can't seem to find in the docs. | 23:35 |
| v1k0d3n | but it's ok...good learning. | 23:35 |
| johnsom | I know kolla has an Octavia module, but I don't know anything about it. Sorry. | 23:37 |
| v1k0d3n | oh that's ok. | 23:39 |
| v1k0d3n | it's fairly basic with the assumption that a user can override the config files at build. i just didn't know exactly what all is setup by default, and what things i needed to override. so it takes a little trial and error. i'm just really happy to have a little direction. this is definitely helping me through it. | 23:40 |
| johnsom | v1k0d3n Here is a tip, you can look at a sample configuration file from our test jobs: http://logs.openstack.org/89/659689/13/check/octavia-v2-dsvm-scenario/8f9d85a/controller/logs/etc/octavia/octavia_conf.txt.gz | 23:42 |
| johnsom | The values aren't always going to be what you want in production, many of the timeouts are very high as the test hosts are virtual-inside-virtual without hardware acceleration. | 23:43 |
| johnsom | But it might be a helpful guide to "what must be set" | 23:43 |
| v1k0d3n | yeah this is really good. | 23:46 |
| v1k0d3n | question, when i create an lb via command-line, and override the network id that's provided in octavia.conf...why would i still get an error that no network has been provided? | 23:47 |
| johnsom | Ah, they are two different network configurations | 23:47 |
| johnsom | The lb-mgmt-net, set in the octavia.conf, is only used for command/control and heartbeats with the Octavia controller processes. The controllers talk to the amphroa, the amphora send heartbeats. No tenant traffic ever goes over this network. | 23:48 |
| johnsom | When you create an LB, you specify your VIP network. This is where tenant traffic will enter the load balancer. It is hot-plugged into a network namespace inside the amphroa where the actual load balancing engine runs. | 23:49 |
| johnsom | So, after you create a load balancer, if you do an "openstack server list", you will see two networks attached to the server. (assuming of course you don't try to put your VIP on your lb-mgmt-net) | 23:50 |
| johnsom | VIP networks can be public or private tenant networks. Basically anything that neutron knows about and the tenant has access to. | 23:50 |
| v1k0d3n | this makes so much more sense now. having that background makes what's written in docs a lot clearer. | 23:51 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!