*** pcaruana has quit IRC | 00:03 | |
*** markvoelker has joined #openstack-ansible | 00:30 | |
*** spatel has joined #openstack-ansible | 00:32 | |
*** rh-jelabarre has quit IRC | 00:43 | |
*** rh-jelabarre has joined #openstack-ansible | 00:44 | |
*** rh-jelabarre has quit IRC | 00:48 | |
*** gouthamr_ has quit IRC | 00:59 | |
*** gouthamr_ has joined #openstack-ansible | 01:05 | |
*** cshen has joined #openstack-ansible | 01:15 | |
*** cshen has quit IRC | 01:20 | |
*** gouthamr_ has quit IRC | 01:25 | |
*** gouthamr_ has joined #openstack-ansible | 01:37 | |
*** d34dh0r53 has quit IRC | 01:38 | |
*** gyee has quit IRC | 01:38 | |
*** d34dh0r53 has joined #openstack-ansible | 01:41 | |
*** markvoelker has quit IRC | 02:14 | |
*** cshen has joined #openstack-ansible | 02:16 | |
*** cshen has quit IRC | 02:20 | |
*** gouthamr_ has quit IRC | 02:52 | |
*** gouthamr_ has joined #openstack-ansible | 02:56 | |
*** gouthamr_ has quit IRC | 03:26 | |
*** gouthamr_ has joined #openstack-ansible | 03:32 | |
nsmeds | Hey y'all, back to bug you with more questions :) I asked this in #openstack-neutron but no replies and this channel usually seems quite active. Is there any strong opinions between deploying OVS+Neutron DVR versus linuxbridge+legacy Neutron? We are currently deployed with linuxbridge (environment was inherited from previous employee), and are investigating options before moving forward. | 03:43 |
---|---|---|
nsmeds | I understand that traffic for floating IPs + between VMs in different subnets will have to be routed via the network hosts with legacy - and IPv6 and SNAT traffic will be routed via network hosts either way. We're considering sticking with legacy just for simplicity sake, but was curious if I'm missing anything vital here. | 03:45 |
nsmeds | The plan would be to simply scale out more network hosts if we start reaching bottlenecks. | 03:45 |
*** gouthamr_ has quit IRC | 04:00 | |
*** markvoelker has joined #openstack-ansible | 04:08 | |
*** gouthamr_ has joined #openstack-ansible | 04:11 | |
*** cshen has joined #openstack-ansible | 04:17 | |
*** cshen has quit IRC | 04:21 | |
*** gouthamr_ has quit IRC | 04:29 | |
*** gouthamr_ has joined #openstack-ansible | 04:30 | |
*** evrardjp has quit IRC | 04:33 | |
*** evrardjp has joined #openstack-ansible | 04:33 | |
*** gouthamr_ has quit IRC | 04:35 | |
*** gouthamr_ has joined #openstack-ansible | 04:40 | |
*** spatel has quit IRC | 04:52 | |
*** gokhani has joined #openstack-ansible | 05:04 | |
*** udesale has joined #openstack-ansible | 05:13 | |
*** markvoelker has quit IRC | 05:13 | |
*** markvoelker has joined #openstack-ansible | 05:14 | |
*** markvoelker has quit IRC | 05:18 | |
noonedeadpunk | jrosser: yep, as it was created in train:) | 05:22 |
noonedeadpunk | So it's already part of T&U codebase | 05:23 |
noonedeadpunk | nsmeds: also there were talks in neutron to obsolete lxb, but dunno if that happen - too many ppl using lxb these days | 05:24 |
noonedeadpunk | and tbh I'd say to consider even ovn instead of ovs (if you're about to choose ovs). But personally I like lxb as well because of it's simplicity as well. | 05:25 |
*** cshen has joined #openstack-ansible | 06:17 | |
*** cshen has quit IRC | 06:21 | |
*** this10nly has joined #openstack-ansible | 06:33 | |
*** pcaruana has joined #openstack-ansible | 06:37 | |
*** CeeMac has joined #openstack-ansible | 06:44 | |
CeeMac | morning | 06:45 |
*** gouthamr_ has quit IRC | 06:50 | |
*** gouthamr_ has joined #openstack-ansible | 06:52 | |
*** arkan has joined #openstack-ansible | 06:54 | |
*** sri_ has quit IRC | 06:54 | |
*** sri_ has joined #openstack-ansible | 06:56 | |
*** arkan has quit IRC | 07:01 | |
*** gouthamr_ has quit IRC | 07:04 | |
*** gouthamr_ has joined #openstack-ansible | 07:05 | |
*** cshen has joined #openstack-ansible | 07:09 | |
*** gouthamr_ has quit IRC | 07:15 | |
*** gouthamr_ has joined #openstack-ansible | 07:22 | |
*** arkan has joined #openstack-ansible | 07:33 | |
*** markvoelker has joined #openstack-ansible | 07:43 | |
*** gouthamr_ has quit IRC | 07:46 | |
*** markvoelker has quit IRC | 07:48 | |
*** gouthamr_ has joined #openstack-ansible | 07:52 | |
*** sshnaidm has joined #openstack-ansible | 07:55 | |
*** gouthamr_ has quit IRC | 07:57 | |
*** udesale_ has joined #openstack-ansible | 08:00 | |
*** udesale has quit IRC | 08:03 | |
*** fyx has quit IRC | 08:04 | |
*** fyx has joined #openstack-ansible | 08:04 | |
*** gouthamr_ has joined #openstack-ansible | 08:09 | |
*** udesale_ has quit IRC | 08:10 | |
arkan | hi guys | 08:22 |
-arkan- jrosser: I've done what you've told me for updating os Horizon and it worked, now I don't receive errors about availability zones | 08:25 | |
arkan | for octavia | 08:25 |
arkan | Now I have issue also about octavia "No route to host", and I observed that it could not reach amphora. I found through traceroute it's taking the wrong path | 08:27 |
-arkan- so I've deleted one route, then it started to reach amphora | 08:27 | |
arkan | see here | 08:27 |
arkan | http://paste.openstack.org/show/795685/ | 08:27 |
*** shyamb has joined #openstack-ansible | 08:31 | |
arkan | And the problem that I have now, when I create a new LB in Horizon, octavia-housekeeping can not reach the new amphora instance | 08:32 |
arkan | # compute network config | 08:38 |
arkan | http://paste.openstack.org/show/795687/ | 08:38 |
arkan | # controller network config | 08:38 |
arkan | http://paste.openstack.org/show/795688/ | 08:38 |
arkan | # openstack_user_config.yml | 08:38 |
arkan | http://paste.openstack.org/show/795689/ | 08:38 |
arkan | # user_variables.yml | 08:38 |
arkan | http://paste.openstack.org/show/795690/ | 08:38 |
arkan | # brctl show on compute | 08:38 |
arkan | http://paste.openstack.org/show/795691/ | 08:38 |
arkan | # brctl show on controller | 08:38 |
arkan | http://paste.openstack.org/show/795692/ | 08:38 |
arkan | everyting is working except Octavia | 08:53 |
arkan | SSH to vms through Floating ips are working, only octavia has problem and I'm stuck | 08:54 |
*** shyamb has quit IRC | 08:58 | |
*** arkan has quit IRC | 09:01 | |
admin0 | morning | 09:06 |
*** shyamb has joined #openstack-ansible | 09:09 | |
*** tosky has joined #openstack-ansible | 09:10 | |
*** udesale has joined #openstack-ansible | 09:12 | |
*** arkan_ has joined #openstack-ansible | 09:13 | |
*** arkan_ is now known as arkan | 09:13 | |
arkan | I was disconnected | 09:13 |
admin0 | when is rc2 coming up and will it have an upgrade path ? | 09:22 |
*** shyam89 has joined #openstack-ansible | 09:30 | |
*** shyamb has quit IRC | 09:30 | |
*** markvoelker has joined #openstack-ansible | 09:44 | |
*** markvoelker has quit IRC | 09:48 | |
*** shyamb has joined #openstack-ansible | 09:55 | |
*** shyam89 has quit IRC | 09:57 | |
*** arkan has quit IRC | 10:00 | |
noonedeadpunk | admin0: as I said yestarday, rc2 will be available once release team will approve https://review.opendev.org/#/c/740078/ | 10:23 |
noonedeadpunk | we don't plan any breaking changes between rc2 and final release, but can't give any guarantees regarding that | 10:23 |
admin0 | :) | 10:23 |
admin0 | i have one situation where i need both ussuri and focal :) | 10:24 |
admin0 | i already tried the rc1 .. it worked fine | 10:24 |
noonedeadpunk | in rc1 there was no focal for sure:) | 10:24 |
admin0 | but the installation went in good :D | 10:24 |
admin0 | it worked fine | 10:24 |
noonedeadpunk | btw I'd say that in rc2 we didn't patch all of the projects for focal as well - only core ones and smth near | 10:25 |
noonedeadpunk | so there still might be issues somewhere with focal | 10:25 |
admin0 | in fact, my test domain is up .. https://mero.cloud/ .. focal all the way, 3 nodes, zerossl cert | 10:25 |
admin0 | was able to do most of my tests | 10:25 |
*** arkan has joined #openstack-ansible | 10:26 | |
admin0 | right now, all I need is nova,neutron, glance( nfs), cinder(lvm) and heat | 10:26 |
noonedeadpunk | I'd say that all should generally work | 10:27 |
admin0 | my only point is .. if i move ahead with rc1 now .. when rc2 comes .. would i be able to safely upgrade it | 10:27 |
admin0 | i would like to assume they would work .. since that is hour good our playbooks are ( and ussuri itself is unchanged) | 10:27 |
admin0 | hour -> how | 10:28 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-openstack_hosts stable/ussuri: Add advanced-virtualization CentOS 8 repo https://review.opendev.org/740228 | 10:29 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts stable/ussuri: Add centos-8 support https://review.opendev.org/740230 | 10:30 |
admin0 | noonedeadpunk, in your opinion given the history, when is your expectation of the rc2 :) | 10:30 |
admin0 | are they quick to act .. within 24 hours .. or takes a week generally etc | 10:30 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client stable/ussuri: Add centos-8 support https://review.opendev.org/740231 | 10:30 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_glance stable/ussuri: Add Centos-8 support https://review.opendev.org/740232 | 10:30 |
noonedeadpunk | admin0: all depends tbh, but I think it will be released this week | 10:31 |
noonedeadpunk | eventually, you can just checkout to the SHA provided in that commit | 10:31 |
noonedeadpunk | it will be equally the same as checkout to the tag | 10:31 |
noonedeadpunk | (except of nasty venv names) | 10:32 |
noonedeadpunk | admin0: so the main problem might be mariadb upgrade | 10:32 |
noonedeadpunk | in terms of rc1->rc2 | 10:33 |
admin0 | if its this week, i will wait | 10:33 |
noonedeadpunk | I can't eventually know for sure, it's just my suggestion:) I can't hurry up them :p | 10:33 |
arkan | guys should I try the method used in https://bugs.launchpad.net/openstack-ansible/+bug/1835855 ? | 10:48 |
openstack | Launchpad bug 1835855 in openstack-ansible "octavia lb-mgmt network is broken on production 3 node deployment" [Undecided,New] | 10:48 |
arkan | he solved that by creating veth pair manually | 10:49 |
arkan | in order for arp to work for octavia | 10:49 |
arkan | I don't know if there is another solution, for this, I can see arp on the other side but it does not know to return | 10:50 |
jrosser | arkan: i am still confused by your setup | 10:55 |
jrosser | how do you connect eth12 of the octavia container on the controller to the octavia neturon network? | 10:55 |
arkan | jrosser: it has vlan 510 which is configured on the router also | 10:57 |
arkan | I tried it with ping | 10:57 |
arkan | and it worked | 10:57 |
jrosser | ok, so how does vlan 510 get to eth12 of the octavia container? | 10:57 |
arkan | after I've deleted the route 172.29.232.0/22 | 10:57 |
noonedeadpunk | ah, spatel is not here... maybe you can ping him later about this? | 10:58 |
arkan | yes, this was my dilemma | 10:58 |
jrosser | it is becasue you do not have a bridge for that on your controller | 10:58 |
arkan | how to reach the container | 10:58 |
arkan | eno1.510 | 10:58 |
noonedeadpunk | oh, ok, not interruption | 10:59 |
jrosser | thats not a bridge, look at the diagram here https://github.com/rcbops/rpc-octavia/blob/master/INSTALLATION.md | 10:59 |
arkan | brq6c4d9df6-50 | 10:59 |
arkan | brq6c4d9df6-508000.36572ce402f3noeno1.510 | 10:59 |
jrosser | that is the neutron network | 10:59 |
arkan | humm | 11:00 |
arkan | - network: | 11:00 |
arkan | container_bridge: "br-ext" | 11:00 |
arkan | container_type: "veth" | 11:00 |
arkan | container_interface: "eth14" | 11:00 |
arkan | host_bind_override: "eno1" | 11:00 |
arkan | type: "vlan" | 11:00 |
arkan | ip_from_q: "lbaas" | 11:00 |
arkan | range: "510:520" | 11:00 |
arkan | net_name: "external" | 11:00 |
arkan | group_binds: | 11:00 |
jrosser | plase use a paste for this :) | 11:00 |
arkan | - neutron_linuxbridge_agent | 11:00 |
arkan | - octavia-worker | 11:00 |
arkan | - octavia-housekeeping | 11:00 |
arkan | - octavia-health-manager | 11:00 |
arkan | I used host bind | 11:00 |
arkan | aha | 11:00 |
arkan | I was in a hurry | 11:00 |
jrosser | yes that is for neutron, it is different | 11:00 |
*** arkan has quit IRC | 11:00 | |
jrosser | neutron creates virtualised networks to connect your virtual machines to, and the octaiva amphora | 11:01 |
*** arkan has joined #openstack-ansible | 11:01 | |
jrosser | neutron does not know anything about OSA containers or openstack control plane | 11:01 |
arkan | back | 11:01 |
jrosser | so that is what "container_bridge" is for | 11:01 |
arkan | I've used br-ext bridge and overwritten by using eno1 | 11:01 |
arkan | because this bridge is untagged | 11:02 |
noonedeadpunk | jrosser: do you think we can do smth to easy life with https://bugs.launchpad.net/openstack-ansible/+bug/1835855 ? I'm not sure I 100% understand the issue here, so... | 11:02 |
openstack | Launchpad bug 1835855 in openstack-ansible "octavia lb-mgmt network is broken on production 3 node deployment" [Undecided,New] | 11:02 |
jrosser | arkan: the traffic is on vlan tag 510 on br-ext though? it's not untagged? | 11:02 |
jrosser | arkan: take the diagram from here https://github.com/rcbops/rpc-octavia/blob/master/INSTALLATION.md | 11:03 |
arkan | yes, but in my network config, I did not attach it to vlan | 11:03 |
noonedeadpunk | maybe we can at least copy https://github.com/rcbops/rpc-octavia/blob/master/INSTALLATION.md to our docs... | 11:03 |
arkan | because in this case it will be tagged 2 times, one by my network config, and one by neutron | 11:03 |
jrosser | yes we should do that | 11:03 |
jrosser | arkan: no, not at all | 11:03 |
jrosser | the diagram explains it all | 11:03 |
jrosser | there are two completely distinct parts of that diagram | 11:04 |
arkan | yes | 11:05 |
jrosser | on the left there is br-lbaas, the veth and v-br-lbaas.111 | 11:05 |
arkan | yes | 11:05 |
jrosser | those are things that the deployer must create on the controller in order for the octavia container to connect onto (in that diagram) vlan 111 | 11:05 |
jrosser | on the right hand side of the diagram is the things that really neutron cares about | 11:05 |
arkan | what about compute node ? | 11:06 |
jrosser | but in your case this is confusing becasue you have got both the octaiva containers *and* the neutron L3 agent running on the same controller host | 11:06 |
jrosser | so you need to satisfy the octavia container being able to connect to br-lbaas (diagram left) | 11:07 |
jrosser | and tell neutron which vlan the octavia mgmt network is on (diagram right) | 11:07 |
arkan | which is lbaas-mgmt, right ? | 11:07 |
jrosser | yes, a logical network in neutron, vlan type | 11:08 |
jrosser | but you cannot connect octavia container to that because neutron does not know anything about that | 11:08 |
jrosser | so you de-tag yourself and make a bridge for the container | 11:08 |
arkan | so we have here 2 bridges, br-lbaas and br-vlan | 11:09 |
arkan | and lbaas-mgmt Neutron network | 11:09 |
arkan | these should go inside controller | 11:09 |
arkan | should I create the same bridges inside compute node ? | 11:09 |
jrosser | yes and you need to take untagged packets from br-lbass and put them on the right tag on the neutron interface | 11:09 |
noonedeadpunk | TIL | 11:10 |
arkan | then it means that I need to use the scripts that are on the page | 11:10 |
jrosser | no, becasue when you create octavia amphora it knows which vlan to use from the physical interface becasue of how you have set up lbaas-mgmt in neutron | 11:10 |
jrosser | you need to do the necessary in your netplan to make a bridge for octavia contaienrs that is not tagged | 11:11 |
jrosser | but this is nothing at all to do with the neutron config | 11:11 |
arkan | I don't know how I can do it on netplan, I exausted the eno1 | 11:12 |
openstackgerrit | Merged openstack/openstack-ansible-os_neutron stable/ussuri: Add centos-8 support https://review.opendev.org/739858 | 11:14 |
jrosser | like you have here http://paste.openstack.org/show/795688/ | 11:14 |
jrosser | just make another after br-vlan, and add eno1.510 to vlans | 11:14 |
*** dave-mccowan has joined #openstack-ansible | 11:18 | |
*** markvoelker has joined #openstack-ansible | 11:19 | |
*** markvoelker has quit IRC | 11:24 | |
arkan | jrosser: these br-vlan are not coneected to the router, eno2 has direct cable from controller to compute node | 11:25 |
*** shyamb has quit IRC | 11:25 | |
jrosser | the router does not matter | 11:25 |
arkan | ok | 11:26 |
jrosser | these networks are isolated from each other, just L2 vlans | 11:26 |
arkan | yes | 11:26 |
arkan | so I will make br-lbaas bridge after br-vlan and give it vlan 510 | 11:27 |
arkan | using eno1.510 | 11:27 |
jrosser | sorry i did not know about eno2 | 11:28 |
jrosser | pick the interface you need | 11:28 |
arkan | I can use eno2.510 | 11:28 |
jrosser | which ever one has the neutron lbaas-mgmt network on <- thats important | 11:28 |
arkan | the controller has, but I will create identical config for br-lbaas on compute node also | 11:29 |
arkan | so both controller and compute nodes will have br-lbaas with interface attached eno2.510 | 11:30 |
jrosser | it is not necessary on the compute node but will not harm | 11:30 |
arkan | in netplan when I attache the interface as in the form eno2.510, it means in the vlan section I will create vlan for it, which in my case has seg. id 510 | 11:31 |
arkan | jrosser: so this is the proposed solution, is to create br-lbaas with attached interface 510 | 11:32 |
arkan | jrosser: so this is the proposed solution, is to create br-lbaas with attached interface eno2.510 | 11:33 |
arkan | I don't need to run some scripts as it's shown in the page to create veth stuff | 11:33 |
*** shyamb has joined #openstack-ansible | 11:33 | |
arkan | jrosser: https://github.com/rcbops/rpc-octavia/blob/master/INSTALLATION.md the bridge has an IP address, should I give IP address also, or it's not important ? | 11:38 |
*** shyam89 has joined #openstack-ansible | 11:50 | |
*** shyamb has quit IRC | 11:53 | |
*** rh-jelabarre has joined #openstack-ansible | 11:57 | |
jrosser | arkan: i do not know why they bridge has an IP address there, i don't think it is necessary | 12:00 |
arkan | jrosser: ok, I was also checking, they are using script | 12:00 |
arkan | Netplan support also script | 12:01 |
arkan | I'm checking the doc | 12:01 |
arkan | in order to connect br-vlan with br-lbaas | 12:01 |
arkan | jrosser: in the script they created veth, and connected it to both br-vlan and br-lbaas | 12:02 |
*** shyam89 has quit IRC | 12:02 | |
arkan | but on the br-lbaas side it has vlan id | 12:02 |
arkan | also the network provider type is not vlan, but it's "raw" for br-lbaas | 12:06 |
jrosser | it is a different environment | 12:07 |
arkan | but the connection veth it needs to be created by me, and make it post-up, for connecting br-lbaas and br-vlan | 12:09 |
arkan | as it's shown in the page | 12:09 |
jrosser | you can connect it wherever is needed | 12:11 |
arkan | yes, but if I forget to do so, then I will have issues with Octava | 12:12 |
arkan | *Octavia | 12:12 |
arkan | I rather to not forget and make Octavia happy | 12:12 |
arkan | in return then I can go to sleep :)) | 12:13 |
jrosser | why not make it just like you have done with br-storage? | 12:13 |
arkan | humm | 12:14 |
jrosser | br-strorage is vlan 20 (untagged) on eno1 (tagged) | 12:14 |
jrosser | this is just the same thing but br-lbaas and vlan 510 | 12:14 |
arkan | it's tagged | 12:14 |
jrosser | well it shouldnt be | 12:15 |
jrosser | becasue br-storage is another interface which is there in order to wire veth up to the control plane containers | 12:15 |
jrosser | it should be untagged on the bridge | 12:15 |
jrosser | it should be tagged on the physical interface | 12:15 |
arkan | yes, it's tagged in the physical interface | 12:15 |
arkan | as in my netplan in the vlans: section, it is attached to eno1.20 | 12:16 |
jrosser | so you can do just the same thing for br-lbaas? | 12:16 |
arkan | sure I can do it | 12:16 |
arkan | will it work like that ? | 12:17 |
arkan | if I use it as I use br-storage, it would be awesome | 12:17 |
arkan | then I will not use neither type vlan nor host_bind_override in the network provider in openstack_user_config.yml | 12:19 |
arkan | as I did for br-storage | 12:19 |
arkan | br-storage uses 'raw' and it does not use host_bind_override | 12:19 |
arkan | brilliant idea | 12:20 |
arkan | I hope this will work | 12:20 |
arkan | humm but wait | 12:21 |
jrosser | that is all neutron config | 12:21 |
arkan | jrosser: I think I tried it someday, if I'm not mistaken | 12:22 |
jrosser | as i said this is not to do with neutron - don't break what you've already got working | 12:22 |
jrosser | remeber the two halves of the diagram | 12:22 |
jrosser | the stuff in provider_networks is giving config for both parts of that | 12:22 |
jrosser | container_<..> for the left part | 12:22 |
jrosser | other fields for neutron | 12:22 |
arkan | jrosser: I need to add a new network: section in "provider_networks" for br-lbaas | 12:23 |
arkan | but just I want to tell you | 12:24 |
arkan | if I use "raw" I think I will get an error | 12:24 |
arkan | telling that "can not enslave bridge ..." | 12:24 |
arkan | I think I was hit with this type of error | 12:24 |
arkan | aha maybe it was with flat type | 12:25 |
arkan | not raw | 12:25 |
arkan | yes, I think it was with "flat" type I got that error | 12:25 |
openstackgerrit | Chandan Kumar (raukadah) proposed openstack/openstack-ansible master: [WIP] improve ironic tempest testing https://review.opendev.org/736507 | 12:32 |
*** stingrayza has joined #openstack-ansible | 12:39 | |
*** also_stingrayza has quit IRC | 12:40 | |
*** arkan has quit IRC | 13:01 | |
*** spatel has joined #openstack-ansible | 13:03 | |
openstackgerrit | Merged openstack/openstack-ansible master: Add Centos-8 support https://review.opendev.org/689629 | 13:12 |
spatel | noonedeadpunk: jrosser oh boy! this is a big win!! merged https://review.opendev.org/689629 & https://review.opendev.org/#/c/735781/ | 13:25 |
spatel | we are inch closer to release centos-8 :) | 13:26 |
noonedeadpunk | spatel: btw about that. can you recheck all our centos8 changes? | 13:33 |
noonedeadpunk | as we have checks now merged - thay will test it as well | 13:33 |
noonedeadpunk | so that we know they're passing CI | 13:34 |
spatel | let me see.. | 13:34 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible stable/ussuri: Add Centos-8 support https://review.opendev.org/740289 | 13:37 |
noonedeadpunk | spatel: I mean only not merged ones | 13:37 |
noonedeadpunk | they're mostly yours :) | 13:37 |
noonedeadpunk | and for master :p | 13:38 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Add corosync/pacemaker installation https://review.opendev.org/739146 | 13:39 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_masakari master: Add CentOS 8 support https://review.opendev.org/739148 | 13:39 |
spatel | noonedeadpunk: i have DM you.. for recheck help | 13:39 |
*** arkan_ has joined #openstack-ansible | 13:47 | |
*** arkan_ is now known as arkan | 13:47 | |
spatel | noonedeadpunk: i did pushed them for recheck, lets see the result in few hour | 13:48 |
*** mindthecap has joined #openstack-ansible | 13:53 | |
noonedeadpunk | cool! many thanks! | 13:56 |
mindthecap | hey! Has someone thought about adding image-volume cache on cinder? I can add it manually atleat for OSA Train but the configuration get overwritten by ansible for obvious reasons. Default values don't intervene with cinder so when cache isn't working, it failovers to downloading image from glance. | 13:59 |
*** arkan has quit IRC | 14:01 | |
*** this10nly has quit IRC | 14:07 | |
*** mindthecap has quit IRC | 14:10 | |
mgariepy | mindthecap, i do use the image_volume_cache via config overrides | 14:10 |
*** mindthecap has joined #openstack-ansible | 14:10 | |
*** mindthecap has quit IRC | 14:10 | |
spatel | noonedeadpunk: they are still failing for multiple reason, may be something odd going on in CI pipeline. | 14:19 |
*** cshen has quit IRC | 14:27 | |
*** chandankumar has joined #openstack-ansible | 14:39 | |
openstackgerrit | Chandan Kumar (raukadah) proposed openstack/openstack-ansible-os_ironic master: [DNM] test patch https://review.opendev.org/727067 | 14:46 |
*** arkan has joined #openstack-ansible | 15:36 | |
*** gyee has joined #openstack-ansible | 15:50 | |
*** udesale_ has joined #openstack-ansible | 16:08 | |
*** udesale has quit IRC | 16:10 | |
*** d34dh0r53 has quit IRC | 16:10 | |
*** udesale_ has quit IRC | 16:27 | |
*** d34dh0r53 has joined #openstack-ansible | 16:30 | |
*** mloza has joined #openstack-ansible | 16:33 | |
openstackgerrit | Merged openstack/openstack-ansible-os_cinder stable/stein: Fix policy configuration https://review.opendev.org/739641 | 16:45 |
*** cshen has joined #openstack-ansible | 16:45 | |
*** mwhahaha_ has joined #openstack-ansible | 16:48 | |
*** mubix_ has joined #openstack-ansible | 16:48 | |
*** dasp_ has joined #openstack-ansible | 16:49 | |
*** cshen has quit IRC | 16:49 | |
*** mwhahaha has quit IRC | 16:55 | |
*** mubix has quit IRC | 16:55 | |
*** nicolasbock has quit IRC | 16:55 | |
*** dasp has quit IRC | 16:55 | |
*** sep has quit IRC | 16:55 | |
*** mubix_ is now known as mubix | 16:55 | |
*** mwhahaha_ is now known as mwhahaha | 16:55 | |
*** fridtjof[m] has quit IRC | 16:56 | |
*** ioni has quit IRC | 16:56 | |
*** masterpe has quit IRC | 16:56 | |
*** nicolasbock has joined #openstack-ansible | 16:58 | |
*** sep has joined #openstack-ansible | 16:58 | |
*** arkan has quit IRC | 17:00 | |
*** ioni has joined #openstack-ansible | 17:03 | |
*** sshnaidm is now known as sshnaidm|afk | 17:12 | |
*** arkan has joined #openstack-ansible | 17:14 | |
arkan | jrosser: I used octavia_provider_network_type: raw and the setup-openstack.yml -> os-octavia-install.yml spit an error " Invalid input for operation: network_type value 'raw' not supported."" | 17:16 |
arkan | this is in provider_networks, http://paste.openstack.org/show/795716/ | 17:17 |
arkan | I wanted to use it as in br-storage, but br-storage does not have - neutron_linuxbridge_agent | 17:18 |
arkan | octavia does not support this type "raw" | 17:18 |
arkan | user_variables.yml ---> http://paste.openstack.org/show/795717/ | 17:19 |
arkan | controller node network config ---> http://paste.openstack.org/show/795718/ | 17:19 |
arkan | Now I re-ran with octavia_provider_network_type: flat, to see if it will work or not | 17:29 |
*** fridtjof[m] has joined #openstack-ansible | 17:32 | |
*** masterpe has joined #openstack-ansible | 17:32 | |
arkan | jrosser: I tried with "flat" and it failed. "physical_network 'lbaas' unknown for flat provider network." | 17:34 |
arkan | so neither "raw" nor "flat" I can use in "octavia_provider_network_type" in the user_variables | 17:34 |
jrosser | you did not need to change any of the neutron stuff | 17:35 |
arkan | you mean I don't need to change anything in openstack_user_config ? | 17:36 |
jrosser | container_bridge was probably the only thing | 17:36 |
arkan | and only to add physical network ? | 17:36 |
arkan | now I've created br-lbaas as a physical network | 17:36 |
jrosser | in neutron? | 17:37 |
-arkan- and I've created provider_network section to use it | 17:37 | |
jrosser | i told you not to touch any of the neutron stuff | 17:37 |
jrosser | we just needed to fix the OSA control plane container being connected to vlan 510, nothing more | 17:37 |
arkan | I've created this here for it http://paste.openstack.org/show/795716/ | 17:38 |
arkan | and now br-ext looks like this after I extracted octavia stuff ---> http://paste.openstack.org/show/795719/ | 17:40 |
arkan | jrosser: the old setup was under the same provider | 17:42 |
arkan | and neutron does not know then about br-lbaas | 17:42 |
jrosser | like i said earlier each provider_networks section can talk about two things, the left and right of the diagram | 17:42 |
jrosser | variables container_<...> are about the control plane | 17:43 |
jrosser | the others are about neutron | 17:43 |
* jrosser heads out | 17:43 | |
arkan | ok, but I did not have any br-lbaas in the network_providers, and the diagram is using it, how neutron will use it if I don't declare it? | 17:44 |
arkan | neutron will connect the left side (container like eth14) ----> bridge (br-lbaas), and my question was how it will do it if I don't declare this? | 17:45 |
*** arkan has quit IRC | 18:00 | |
*** arkan_ has joined #openstack-ansible | 18:01 | |
openstackgerrit | Ryan Drew proposed openstack/openstack-ansible master: Fix KeyError raised when max hostname length exceeded https://review.opendev.org/740343 | 18:01 |
*** arkan_ is now known as arkan | 18:01 | |
*** cshen has joined #openstack-ansible | 18:45 | |
*** cshen has quit IRC | 18:50 | |
*** tosky has quit IRC | 19:26 | |
*** arkan has quit IRC | 20:01 | |
*** arkan has joined #openstack-ansible | 20:01 | |
openstackgerrit | Ryan Drew proposed openstack/openstack-ansible master: Change haproxy galera backend to balance between all galera nodes https://review.opendev.org/740355 | 20:08 |
*** arkan has quit IRC | 20:17 | |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-openstack_hosts master: Replace yum config manager with dnf alternative https://review.opendev.org/739554 | 20:32 |
*** dirk has quit IRC | 20:42 | |
*** cshen has joined #openstack-ansible | 20:45 | |
*** cshen has quit IRC | 20:50 | |
*** jbadiapa has quit IRC | 20:58 | |
*** dirk has joined #openstack-ansible | 21:01 | |
*** yolanda has quit IRC | 21:16 | |
*** yolanda has joined #openstack-ansible | 21:17 | |
*** spatel has quit IRC | 21:52 | |
*** rh-jelabarre has quit IRC | 22:30 | |
*** jamesdenton has quit IRC | 22:40 | |
*** cshen has joined #openstack-ansible | 22:46 | |
*** cshen has quit IRC | 22:50 | |
*** cshen has joined #openstack-ansible | 23:26 | |
*** cshen has quit IRC | 23:30 | |
*** gyee has quit IRC | 23:56 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!