| prometheanfire | cloudnull: do you create netdev files for things like eth0? | 00:01 |
|---|---|---|
| prometheanfire | or rely on the system to do so for you | 00:01 |
| cloudnull | nope no need. physical interfaces should already be present. | 00:02 |
| prometheanfire | ah, guess you wouldn't | 00:02 |
| cloudnull | you can create netdev for physical devices to change the properties of the device itself | 00:02 |
| cloudnull | but that's not required. | 00:02 |
| *** lhinds has quit IRC | 00:02 | |
| prometheanfire | I'm tempted to just go ahead and do that | 00:03 |
| *** portante has quit IRC | 00:03 | |
| *** pabelanger has quit IRC | 00:04 | |
| *** pabelanger has joined #openstack-ansible | 00:05 | |
| *** lhinds has joined #openstack-ansible | 00:06 | |
| *** portante has joined #openstack-ansible | 00:06 | |
| *** masber has joined #openstack-ansible | 00:19 | |
| *** masber has quit IRC | 00:23 | |
| *** masber has joined #openstack-ansible | 00:23 | |
| *** dave-mccowan has joined #openstack-ansible | 00:41 | |
| logan- | quick question for you on https://review.openstack.org/#/c/554389/3 cloudnull | 00:45 |
| cloudnull | answered | 00:47 |
| cloudnull | logan-: mind having a look at https://review.openstack.org/#/c/555910 if you have a free moment | 00:48 |
| logan- | yup will do | 00:49 |
| cloudnull | also http://logs.openstack.org/79/466179/24/check/openstack-ansible-upgrade-ubuntu-xenial/143b1de/job-output.txt.gz#_2018-03-24_00_44_36_740767 -- is this something we're aware of | 00:49 |
| cloudnull | looks like cell mapping is failing in nova? | 00:50 |
| cloudnull | for the upgrade job that is | 00:50 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_barbican master: Convert role to use a common systemd service role https://review.openstack.org/552071 | 00:54 |
| logan- | have not seen that yet | 00:55 |
| logan- | do you think it is something base image related like keystone was? | 00:55 |
| *** odyssey4me has quit IRC | 00:55 | |
| *** odyssey4me has joined #openstack-ansible | 00:56 | |
| logan- | well nevermind | 00:57 |
| logan- | just read thru the traceback | 00:57 |
| logan- | ""Unknown column 'disabled' in 'field list'\"" | 00:57 |
| logan- | https://github.com/openstack/openstack-ansible/blob/62bbe0f38e04864e3e669fb0871ddacefca79814/playbooks/defaults/repo_packages/openstack_services.yml#L167-L169 | 00:58 |
| logan- | wow none of the shas have been bumped in master since 2/16? | 00:58 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Add container journal linking https://review.openstack.org/553707 | 01:09 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Add container journal linking https://review.openstack.org/553707 | 01:10 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Allow deployers to define the container type https://review.openstack.org/554389 | 01:18 |
| cloudnull | I think the role tests are using the head of master. | 01:19 |
| cloudnull | while the integrated gate is behind | 01:19 |
| cloudnull | unless the tests repo is reading that in somewhere | 01:20 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-openstack_hosts master: Revert "Ensure the /etc/pki/rpm-gpg directory exists" https://review.openstack.org/555845 | 01:20 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: Add tests for MultiStringOp https://review.openstack.org/555086 | 01:21 |
| openstackgerrit | Merged openstack/ansible-role-systemd_mount master: Add basic tests https://review.openstack.org/555910 | 01:21 |
| logan- | oh that log is a role test? gotcha | 01:21 |
| logan- | some stuff related to this field just merged in the last day or so https://review.openstack.org/#/q/project:openstack/nova+branch:master+topic:bp/cell-disable | 01:23 |
| logan- | https://github.com/openstack/nova/commit/9377127eb887d263ba1c80990f93329fda3c19bc and https://github.com/openstack/nova/commit/a26ad9f9a82dfb5210e4a979e67e000989d7ff91 etc | 01:24 |
| logan- | and still 3-4 patches still open for that bp | 01:25 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Add container journal linking https://review.openstack.org/553707 | 01:28 |
| *** flemingo_ has quit IRC | 01:28 | |
| cloudnull | logan-: :( | 01:28 |
| cloudnull | we might need to pin the nova role test. | 01:28 |
| cloudnull | or just let them finish working on it | 01:28 |
| logan- | yea idk if it is only affecting upgrades and the regular tests are passing maybe its something worth pinging them about in case its not something addressed by the pending patches | 01:30 |
| logan- | or.. you know.. we could do a depends-on with all of the patches still open | 01:30 |
| logan- | and see if it passes | 01:30 |
| cloudnull | that might be worth a try | 01:31 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: Add tests for MultiStringOp https://review.openstack.org/555086 | 01:39 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts master: Copy the "locale" into base container image https://review.openstack.org/555971 | 01:43 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Add container journal linking https://review.openstack.org/553707 | 01:45 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_barbican master: Convert role to use a common systemd service role https://review.openstack.org/552071 | 01:47 |
| cloudnull | so with the linked journals we can do things like `journalctl --follow --file=/var/log/journal/$(lxc-attach -n aio1_utility_container-21e9790d -- cat /etc/machine-id)/system.journal` on the host which how i found the locales issue | 01:50 |
| cloudnull | https://review.openstack.org/555971 | 01:50 |
| cloudnull | but we can also use journald to ship the journals elsewhere. | 01:51 |
| cloudnull | like https://cloudnull.io/2018/02/journaling-remotely/ | 01:52 |
| * cloudnull shameless plug | 01:52 | |
| *** fatdragon has quit IRC | 01:54 | |
| logan- | :) | 01:55 |
| *** ryan_ has quit IRC | 02:07 | |
| openstackgerrit | Merged openstack/openstack-ansible-plugins master: Add tests for MultiStringOp https://review.openstack.org/555086 | 02:27 |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible-ops master: Adding mnaio qemu file backends https://review.openstack.org/549097 | 02:29 |
| *** dave-mccowan has quit IRC | 03:08 | |
| *** masuberu has joined #openstack-ansible | 03:09 | |
| *** masuberu has quit IRC | 03:11 | |
| *** masber has quit IRC | 03:11 | |
| *** masuberu has joined #openstack-ansible | 03:11 | |
| *** masuberu has quit IRC | 03:14 | |
| *** masuberu has joined #openstack-ansible | 03:15 | |
| *** savvas has joined #openstack-ansible | 03:16 | |
| *** savvas has joined #openstack-ansible | 03:17 | |
| *** cjloader has joined #openstack-ansible | 03:29 | |
| *** cjloader has quit IRC | 03:30 | |
| *** cjloader has joined #openstack-ansible | 03:31 | |
| *** savvas_ has joined #openstack-ansible | 03:34 | |
| *** savvas has quit IRC | 03:35 | |
| *** masuberu has quit IRC | 03:39 | |
| *** savvas has joined #openstack-ansible | 03:56 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-os_keystone master: Re-cache venv if extracted venv is missing https://review.openstack.org/555981 | 03:56 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add playbook to ship journals from hosts https://review.openstack.org/555982 | 03:59 |
| *** savvas has quit IRC | 04:00 | |
| *** cjloader has quit IRC | 04:01 | |
| *** cjloader has joined #openstack-ansible | 04:01 | |
| *** cjloader has quit IRC | 04:02 | |
| *** cjloader has joined #openstack-ansible | 04:03 | |
| *** cjloader has quit IRC | 04:13 | |
| *** cjloader has joined #openstack-ansible | 04:13 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-os_keystone master: Rebuild venv if missing https://review.openstack.org/555981 | 04:32 |
| *** djdule has joined #openstack-ansible | 04:33 | |
| djdule | Hello | 04:33 |
| djdule | I need some help with a setup | 04:33 |
| djdule | I do have OpenStack installed with Ansible of course, latest version, Pike | 04:34 |
| djdule | And everything works except connectivity from cloud to rest of the world | 04:34 |
| *** cjloader has quit IRC | 04:34 | |
| djdule | I created external provider network | 04:34 |
| djdule | but it is still not working. And I am stuck now, no idea what went wrong | 04:35 |
| *** cjloader has joined #openstack-ansible | 04:35 | |
| *** cjloader has quit IRC | 04:37 | |
| *** cjloader has joined #openstack-ansible | 04:37 | |
| *** cjloader has quit IRC | 04:38 | |
| *** cjloader has joined #openstack-ansible | 04:39 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add playbook to ship journals from hosts https://review.openstack.org/555982 | 04:39 |
| cloudnull | djdule: I assume the external router has a gateway on it ? | 04:40 |
| djdule | You mean external router on OpenStack ? | 04:40 |
| cloudnull | and that the network being used as the gateway is routable ? | 04:40 |
| cloudnull | yes, you'll have two network interfaces on your router. one being the gateway. | 04:41 |
| cloudnull | the other being your tenant network | 04:41 |
| djdule | Yes, that is correct | 04:42 |
| cloudnull | in most cases the first network (GATEWAY_NET) is has something off when it was created. | 04:42 |
| *** cjloader has quit IRC | 04:42 | |
| djdule | What do you mean by "first network (GATEWAY_NET) is has something off when it was created" ? | 04:44 |
| cloudnull | this is an old post but still relevant https://developer.rackspace.com/blog/neutron-networking-l3-agent | 04:44 |
| cloudnull | there are a couple flag you have to pass in when the network was created | 04:45 |
| cloudnull | 's/router/network/ | 04:45 |
| djdule | i created network using command: neutron net-create external-net --router:external --provider:physical_network flat --provider:network_type flat | 04:45 |
| cloudnull | ok. | 04:45 |
| djdule | and then subnet neutron subnet-create --name external-subnet --allocation-pool start=139.25.25.200,end=139.25.25.254 --disable-dhcp --gateway 139.25.25.193 external-net 139.25.25.192/26 | 04:45 |
| cloudnull | ok. | 04:46 |
| cloudnull | and that gateway is routable ? | 04:46 |
| cloudnull | can you ping that address within the instance? | 04:46 |
| cloudnull | and are the instances getting DHCP | 04:46 |
| djdule | I can ping and connect to gateway 139.25.25.193 from controller, but not from instance | 04:47 |
| djdule | Yes, instances get DHCP | 04:47 |
| djdule | and I can also associate floating IPs to them | 04:47 |
| djdule | instance to instance works | 04:47 |
| cloudnull | great! | 04:47 |
| djdule | only thing that is not working is connecting to rest of the world from instance | 04:47 |
| cloudnull | is this an AIO | 04:47 |
| djdule | no, I do have separate cloud controller and compute nodes | 04:48 |
| cloudnull | ok. | 04:48 |
| djdule | one cloud controller, not HA | 04:48 |
| cloudnull | cool, that'll work fine. | 04:48 |
| cloudnull | can you assign a float to more than one vm and ping the floats from the vms ? | 04:49 |
| djdule | yes, and I can also SSH to float, tried that already | 04:49 |
| cloudnull | cool! | 04:49 |
| cloudnull | sounds like we're really close | 04:50 |
| djdule | I really hope so, since I already wasted 2 days trying to resolve it :) | 04:50 |
| cloudnull | what's the default gateway on the VM | 04:50 |
| cloudnull | and do you have a gateway set on the router? | 04:50 |
| cloudnull | neutron router-list , router-show $UUID | 04:51 |
| djdule | i have private net 192.168.2.0/24, and default gateway is 192.168.2.1 from within the VM | 04:51 |
| cloudnull | and you can ping 192.168.2.1 | 04:51 |
| djdule | routes | {"destination": "139.25.25.192/26", "nexthop": "139.25.25.193"} | 04:52 |
| djdule | let me check ping | 04:52 |
| cloudnull | ok. | 04:53 |
| cloudnull | I assume you ran "neutron router-gateway-set ..." | 04:53 |
| djdule | I am not 100% sure | 04:54 |
| djdule | what should be whole command, just to check | 04:54 |
| cloudnull | `neutron router-gateway-set $ROUTER_NAME $NET_NAME` | 04:54 |
| djdule | $ROUTER_NAME is my router, $NET_NAME is external network ? | 04:55 |
| djdule | let me try it | 04:55 |
| cloudnull | yes | 04:55 |
| cloudnull | also `neutron router-interface-add $ROUTER_UUID $NETWORK_SUBNET_UUID` | 04:56 |
| * cloudnull is tired and going from memory | 04:56 | |
| cloudnull | but those should be close to the commands you need to run to setup a router once the router is created | 04:57 |
| cloudnull | i gotta run . | 04:59 |
| cloudnull | let us know how it goes | 04:59 |
| cloudnull | hope you make it all work soon djdule | 04:59 |
| djdule | thanks a lot ! | 05:00 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 05:02 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 05:06 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 05:08 |
| *** djdule has quit IRC | 05:13 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 05:20 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 05:32 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 05:50 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_cinder master: Followning the same logic as https://review.openstack.org/#/c/555981/ Making the respective changes to the os_cinder role. https://review.openstack.org/555999 | 06:29 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add playbook to ship journals from hosts https://review.openstack.org/555982 | 06:32 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_barbican master: Convert role to use a common systemd service role https://review.openstack.org/552071 | 06:35 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: [WIP] Disable global rsyslog https://review.openstack.org/556002 | 06:53 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_nova master: Propagate https://review.openstack.org/#/c/555981/ https://review.openstack.org/556003 | 06:57 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_glance master: Propagate https://review.openstack.org/#/c/555981/ https://review.openstack.org/556005 | 07:16 |
| *** djdule has joined #openstack-ansible | 07:22 | |
| djdule | Hi, I have a question about setting up external connectivity | 07:23 |
| djdule | I do have Pike, latest version of openstack-ansible | 07:23 |
| djdule | but I do not understand how to setup mapping between external network and interface | 07:24 |
| djdule | as far as I understand provider network interface should be defined in neutron ml2_conf.ini | 07:24 |
| djdule | but in mine, that is generated by Ansible, I do have | 07:25 |
| djdule | [ml2_type_flat] flat_networks = flat | 07:25 |
| djdule | I do no not understand how that maps to my physical network interface | 07:25 |
| djdule | because when I create net | 07:25 |
| djdule | openstack network create --external --share --provider-physical-network flat --provider-network-type flat provider1 | 07:25 |
| djdule | and all other steps | 07:25 |
| djdule | traffic is not getting out | 07:26 |
| djdule | so it seems to me there is no mapping between network in openstack and my physical interface | 07:26 |
| djdule | any idea ? | 07:26 |
| *** jwitko has quit IRC | 07:27 | |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_neutron master: Propagate https://review.openstack.org/#/c/555981/ https://review.openstack.org/556013 | 07:31 |
| *** chhagarw has joined #openstack-ansible | 07:31 | |
| djdule | any help please ? | 07:33 |
| xdfil | djdule: bridge_mappings = | 07:38 |
| djdule | from what file ? | 07:38 |
| djdule | if you ask about ml2_conf.ini, no such line | 07:39 |
| djdule | that file is auto generated by ansible | 07:39 |
| djdule | openstack-ansible, to be precise | 07:39 |
| xdfil | djdule: bridge_mappings = <name_of network_you_create>:<physical_nic_on_host> | 07:39 |
| xdfil | so like vlan:bond0 | 07:40 |
| xdfil | or vlan:eth0 | 07:40 |
| xdfil | and yeah it gets generated by OSA | 07:40 |
| djdule | but in what file I shall look for that line ? | 07:40 |
| xdfil | i think ml2 | 07:41 |
| xdfil | but one sec | 07:41 |
| djdule | no such line there | 07:41 |
| xdfil | https://docs.openstack.org/project-deploy-guide/openstack-ansible/pike/app-config-test.html | 07:42 |
| xdfil | net-name | 07:42 |
| xdfil | net-name is the first part of the mapping | 07:42 |
| xdfil | and the bridge interface is usually the second part | 07:43 |
| xdfil | depends on how you set it up but thats the easiest way (assuming you created linux bridges on your hosts manually) | 07:43 |
| djdule | ok this is how my deployment file looks like, specifically that part related to bridge | 07:44 |
| djdule | - network: container_bridge: "br-vlan" container_type: "veth" container_interface: "eth12" host_bind_override: "eth12" type: "flat" net_name: "flat" group_binds: - neutron_linuxbridge_agent | 07:44 |
| djdule | so it is exactly the same as example you pointed out | 07:45 |
| djdule | so if I understand properly, that should map flat network to eth12 ? | 07:45 |
| djdule | but it is not reflected in ml2 config file | 07:46 |
| xdfil | I personally dont use flat | 07:46 |
| *** gkadam_ has joined #openstack-ansible | 07:46 | |
| xdfil | so like | 07:46 |
| xdfil | it could be the linux bridge agent conf | 07:47 |
| xdfil | ~ /etc/neutron/plugins/ml2/linux-bri*~~ | 07:47 |
| djdule | ok, but that file does not exist on my deployment | 07:48 |
| xdfil | what host are you on? | 07:48 |
| djdule | on controller | 07:49 |
| xdfil | is it an AIO? | 07:49 |
| djdule | no, not AIO | 07:49 |
| djdule | separated one controller, multiple compute nodes | 07:49 |
| xdfil | neutron_agents container? | 07:49 |
| djdule | ah, sorry, there it is | 07:50 |
| xdfil | actually it might be on neutron_server also | 07:50 |
| djdule | physical_interface_mappings = flat:eth12,vlan:eth11 | 07:50 |
| djdule | it is not on neutron server, it is in agents container | 07:50 |
| xdfil | the agents container is where all the magic happens | 07:50 |
| djdule | ok, there is that line | 07:51 |
| xdfil | although in Queens i think they got rid of it | 07:51 |
| djdule | and if I understand it properly it is saying that flat network is mapped to eth12 | 07:51 |
| xdfil | yes | 07:51 |
| xdfil | inside the container | 07:51 |
| djdule | there is eth12 | 07:51 |
| djdule | created from /etc/network/interfaces | 07:52 |
| djdule | on controller node | 07:52 |
| djdule | it is veth par leading to br-vlan | 07:52 |
| xdfil | yep ^^ that sounds good | 07:52 |
| djdule | br-vlan is created manually, as per OSA documentation | 07:52 |
| djdule | but when I create networks and everything, I cannot go out from OpenStack | 07:53 |
| xdfil | you configure security group? | 07:53 |
| djdule | which, for VMs ? | 07:53 |
| djdule | yes, VM connectivity works inside cloud | 07:53 |
| djdule | and floating IPs | 07:54 |
| xdfil | 99% of all network problems in openstack are because you forgot to set the security group | 07:54 |
| djdule | but from qrouter name space I cannot ping gateway | 07:54 |
| *** gkadam has joined #openstack-ansible | 07:54 | |
| xdfil | inside cant hit google.. ok | 07:54 |
| xdfil | is inside RFC1918? | 07:55 |
| djdule | so I do have security group that allows ICMP and ping | 07:55 |
| djdule | yes, inside is 192.168.2.10..... | 07:55 |
| xdfil | yeah outbound you should be fine | 07:55 |
| djdule | there is gateway, router, etc, | 07:55 |
| djdule | and everything looks ok | 07:55 |
| djdule | and internaly it works | 07:55 |
| djdule | i can go between VMs using internal and floating IPs | 07:56 |
| djdule | but I cannot go out | 07:56 |
| djdule | so I am looking into qrouter name space on controller, inside agent container | 07:56 |
| xdfil | so SNAT might be disabled on that subnet | 07:56 |
| *** gkadam_ has quit IRC | 07:56 | |
| xdfil | but by default it should be on | 07:57 |
| xdfil | pretty sure | 07:57 |
| djdule | it is on | 07:57 |
| djdule | root@infra1-neutron-agents-container-e800e983:/# ip netns list qrouter-41eae7b4-333b-4a46-babb-bf7b713b95e1 (id: 1) qdhcp-9757992a-c357-49c9-9d49-1e7fbf87aa3f (id: 2) r | 07:57 |
| djdule | so docs says that I should be able to ping my gateway from qrouter namespace | 07:57 |
| djdule | but I cannot | 07:57 |
| xdfil | yeah you should... | 07:58 |
| xdfil | this sounds familiar | 07:58 |
| xdfil | you have l3 agents running where? | 08:00 |
| djdule | in container on controller node | 08:01 |
| xdfil | what about computes? | 08:01 |
| djdule | they are on separate servers | 08:01 |
| xdfil | not saying you should, but is the neutron-l3 agent running on computes? | 08:01 |
| djdule | let me check, but it should | 08:02 |
| xdfil | and is neutron linux-bridge agent running everywhere | 08:02 |
| djdule | linux-bridge agent is running, ye | 08:03 |
| djdule | l3 no | 08:03 |
| djdule | l3 not on compute node | 08:03 |
| xdfil | ok thats probably fine | 08:03 |
| xdfil | sorry i'm temporarily out of ideas | 08:05 |
| djdule | me too, and I am scratching my head for 2 days | 08:06 |
| xdfil | i should have been in bed hours ago | 08:06 |
| xdfil | hehe ya | 08:06 |
| djdule | good night and thanks :) | 08:06 |
| xdfil | been there | 08:06 |
| xdfil | you'll get it you seem to have a good grasp on these things | 08:06 |
| *** hamza21 has joined #openstack-ansible | 08:13 | |
| *** chhagarw has quit IRC | 08:16 | |
| xdfil | djdule: hey | 08:17 |
| djdule | yes | 08:17 |
| xdfil | get rid of that host_bind_override: "eth12" | 08:17 |
| djdule | and rerun which part ? | 08:17 |
| djdule | when I say rerun, what playbooks to re-apply ? | 08:18 |
| xdfil | the key in neutron linux bridge conf should be "bridge mappings i think | 08:18 |
| xdfil | um... lxc stuff i think | 08:18 |
| xdfil | and then neutron | 08:19 |
| xdfil | so i think the reson you have physical-interface-mappings instead is because of that bind-override | 08:19 |
| djdule | ok, let me try and see | 08:20 |
| xdfil | which you don't need probably and " eth12 " is certainly not the name of the physical interface | 08:20 |
| xdfil | do you have "br-vlan" configured on the compute? | 08:20 |
| *** spinoshi has joined #openstack-ansible | 08:20 | |
| djdule | yes, there is br-vlan on compute | 08:21 |
| xdfil | if that works, thank Pablo Escobar | 08:21 |
| xdfil | ok good | 08:21 |
| xdfil | watching him on TV and that thought just came to me lol | 08:22 |
| djdule | LOL | 08:22 |
| xdfil | he's so determined. its inspiring | 08:22 |
| djdule | ok, reapplying stuff now, let's see if it will help | 08:24 |
| djdule | it will take some time I think | 08:25 |
| djdule | so I will go to sort out some other things and let you know later how it went | 08:25 |
| djdule | but thanks for help | 08:25 |
| xdfil | one thing | 08:25 |
| xdfil | physical_interface_mappings | 08:25 |
| xdfil | what does it say on your compute host | 08:25 |
| xdfil | linux-bridge-agent conf | 08:26 |
| djdule | let me see | 08:26 |
| xdfil | or does it say bridge_mappings | 08:26 |
| djdule | physical_interface_mappings = flat:eth12,vlan:eth11 | 08:26 |
| xdfil | theres your problem | 08:26 |
| xdfil | the plays should fix that | 08:27 |
| *** spinoshi has quit IRC | 08:27 | |
| djdule | what it should say ? | 08:27 |
| xdfil | it should say br-vlan | 08:27 |
| djdule | ok, great, thanks. let me re run stuff and see if it performs | 08:28 |
| xdfil | nn | 08:28 |
| *** djdule has quit IRC | 08:34 | |
| *** sep has quit IRC | 09:15 | |
| *** sep has joined #openstack-ansible | 09:15 | |
| *** yolanda has quit IRC | 09:43 | |
| *** masber has joined #openstack-ansible | 09:50 | |
| *** yolanda has joined #openstack-ansible | 10:00 | |
| *** chhagarw has joined #openstack-ansible | 10:30 | |
| *** sep has quit IRC | 11:19 | |
| *** sep has joined #openstack-ansible | 11:20 | |
| *** masber is now known as masuberu | 11:32 | |
| *** masuberu is now known as masber | 11:35 | |
| *** masber has quit IRC | 11:41 | |
| *** masber has joined #openstack-ansible | 11:42 | |
| *** masber has quit IRC | 11:45 | |
| *** masber has joined #openstack-ansible | 11:47 | |
| *** odyssey4me has quit IRC | 12:07 | |
| *** odyssey4me has joined #openstack-ansible | 12:08 | |
| *** init has joined #openstack-ansible | 12:19 | |
| *** Kande has joined #openstack-ansible | 12:26 | |
| Kande | Hello! | 12:26 |
| *** savvas has joined #openstack-ansible | 12:32 | |
| *** savvas has joined #openstack-ansible | 12:32 | |
| Kande | Hello! Yesterday i exectuded 3 playbooks. In the yml file (openstack_user_config.yml) i defined cidr_networks: container: 172.18.1.0/24 tunnel: 10.0.1.0/24 storage: 10.0.2.0/24 used_ips: - "172.18.1.99,172.18.1.254" - "10.0.1.100,10.0.1.254" - "10.0.2.1,10.0.2.10" - "10.0.2.100,10.0.2.254". But somehow if I execute lxc-info <container> the ip is 10.0.3.0/24. Am I missing something? | 12:38 |
| *** savvas_ has joined #openstack-ansible | 12:38 | |
| *** chhagarw has quit IRC | 12:40 | |
| *** savvas_ has quit IRC | 12:40 | |
| *** savvas has quit IRC | 12:42 | |
| *** radeks has joined #openstack-ansible | 12:52 | |
| *** 07IACPO2U has joined #openstack-ansible | 12:58 | |
| *** dave-mccowan has joined #openstack-ansible | 13:10 | |
| *** dave-mcc_ has joined #openstack-ansible | 13:14 | |
| *** dave-mccowan has quit IRC | 13:16 | |
| *** madhu has quit IRC | 13:20 | |
| *** init has quit IRC | 13:20 | |
| *** Kande has quit IRC | 13:21 | |
| *** nyloc has quit IRC | 13:25 | |
| *** nyloc has joined #openstack-ansible | 13:27 | |
| *** radeks has quit IRC | 13:42 | |
| *** savvas has joined #openstack-ansible | 14:15 | |
| *** savvas has quit IRC | 14:22 | |
| *** savvas has joined #openstack-ansible | 14:22 | |
| *** savvas_ has joined #openstack-ansible | 14:32 | |
| d34dh0r53 | Kande: The 10.0.3.x network is the LXC internal network (https://docs.openstack.org/project-deploy-guide/openstack-ansible/pike/overview-network-arch.html) the networks defined in openstack_user_config are attached to the br-{mgmt, storage, vxlan, vlan} networks that openstack uses | 14:34 |
| *** savvas has quit IRC | 14:35 | |
| *** savvas has joined #openstack-ansible | 14:37 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 14:38 |
| *** savvas_ has quit IRC | 14:41 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 14:41 |
| openstackgerrit | Doug Hellmann proposed openstack/openstack-ansible master: add lower-constraints job https://review.openstack.org/556026 | 14:45 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: Copy the "locale" into base container image https://review.openstack.org/555971 | 14:46 |
| openstackgerrit | Doug Hellmann proposed openstack/openstack-ansible-apt_package_pinning master: add lower-constraints job https://review.openstack.org/556027 | 14:46 |
| openstackgerrit | Doug Hellmann proposed openstack/openstack-ansible-ceph_client master: add lower-constraints job https://review.openstack.org/556028 | 14:46 |
| openstackgerrit | Doug Hellmann proposed openstack/openstack-ansible-galera_client master: add lower-constraints job https://review.openstack.org/556029 | 14:46 |
| openstackgerrit | Doug Hellmann proposed openstack/openstack-ansible-galera_server master: add lower-constraints job https://review.openstack.org/556030 | 14:46 |
| openstackgerrit | Doug Hellmann proposed openstack/openstack-ansible-haproxy_server master: add lower-constraints job https://review.openstack.org/556031 | 14:46 |
| openstackgerrit | Doug Hellmann proposed openstack/openstack-ansible-lxc_container_create master: add lower-constraints job https://review.openstack.org/556032 | 14:47 |
| openstackgerrit | Doug Hellmann proposed openstack/openstack-ansible-lxc_hosts master: add lower-constraints job https://review.openstack.org/556033 | 14:47 |
| openstackgerrit | Doug Hellmann proposed openstack/openstack-ansible-memcached_server master: add lower-constraints job https://review.openstack.org/556034 | 14:47 |
| *** savvas has quit IRC | 14:48 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create master: Allow deployers to define the container type https://review.openstack.org/554389 | 14:52 |
| *** 07IACPO2U has quit IRC | 14:53 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 14:54 |
| *** savvas has joined #openstack-ansible | 14:56 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create master: Add container journal linking https://review.openstack.org/553707 | 14:58 |
| *** savvas has quit IRC | 15:00 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 15:06 |
| *** savvas has joined #openstack-ansible | 15:09 | |
| *** savvas has quit IRC | 15:14 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 15:14 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder master: Execute cinder_backends on the last API server https://review.openstack.org/556046 | 15:15 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder master: Execute cinder_backends on the last API server https://review.openstack.org/556046 | 15:20 |
| odyssey4me | evrardjp jmccrory hwoarang logan- I think https://review.openstack.org/556046 may be the simplest way of solving that cinder API LB issue. | 15:23 |
| *** radeks has joined #openstack-ansible | 15:31 | |
| *** masber has quit IRC | 15:32 | |
| *** masuberu has joined #openstack-ansible | 15:32 | |
| *** savvas has joined #openstack-ansible | 15:36 | |
| *** savvas has quit IRC | 15:41 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 15:44 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 15:53 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 16:03 |
| *** sawblade_ has quit IRC | 16:08 | |
| evrardjp | hey odyssey4me | 16:12 |
| evrardjp | how are you today? | 16:12 |
| evrardjp | I will check very quick | 16:12 |
| evrardjp | odyssey4me: it looks good to me indeed | 16:14 |
| evrardjp | depends on the rise time though | 16:14 |
| logan- | there are a lot of retries on the task | 16:15 |
| logan- | so hopefully that wont be an issue | 16:15 |
| logan- | also assuming a 3 node infra you get 2 role runs (api1 and then api2) for the endpoint to come up | 16:18 |
| *** openstackgerrit has quit IRC | 16:18 | |
| evrardjp | 2 role runs? | 16:18 |
| evrardjp | You mean the serialization steps in? | 16:18 |
| evrardjp | not so sure to understand | 16:19 |
| logan- | yeah it is in serial, so it deploys api1, brings api1 endpoint up, deploys api2, api2 endpoint up, and then api3 runs and you get most of api3's run before it will ever try to use the LB endpoint | 16:19 |
| evrardjp | I think it's worth merging if it's not breaking anything, it can't do worse than what it is right now | 16:19 |
| evrardjp | yeah if it's serialized enough we should be fine | 16:19 |
| logan- | ++ hope so | 16:20 |
| evrardjp | but the serialization should have the same order as the inventory, which is the case by default | 16:20 |
| evrardjp | it's just a more hidden thing :) | 16:20 |
| evrardjp | I voted positively, I have to go for now | 16:21 |
| evrardjp | I've checked experimental | 16:21 |
| evrardjp | if it's not merged when I come back and the integrated role passes, I'll merge itby then | 16:21 |
| evrardjp | ttyl | 16:22 |
| logan- | ok.just fyi experimental won't tell us anything about the fix because all lb management tasks skip in the gate due to only 1 cinder-api being present | 16:22 |
| *** savvas has joined #openstack-ansible | 16:26 | |
| *** savvas has quit IRC | 16:30 | |
| *** openstackgerrit has joined #openstack-ansible | 16:39 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 16:39 |
| *** savvas has joined #openstack-ansible | 16:43 | |
| *** cjloader has joined #openstack-ansible | 16:45 | |
| *** cjloader has quit IRC | 16:46 | |
| *** cjloader has joined #openstack-ansible | 16:47 | |
| *** cjloader has quit IRC | 16:52 | |
| *** cjloader has joined #openstack-ansible | 16:53 | |
| *** savvas has quit IRC | 17:00 | |
| *** savvas has joined #openstack-ansible | 17:00 | |
| *** gkadam has quit IRC | 17:01 | |
| *** savvas has quit IRC | 17:07 | |
| *** savvas has joined #openstack-ansible | 17:07 | |
| *** savvas has quit IRC | 17:22 | |
| *** cjloader has quit IRC | 17:24 | |
| *** radeks_ has joined #openstack-ansible | 17:27 | |
| *** radeks has quit IRC | 17:27 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 17:31 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: Only su within container if container_user is provided https://review.openstack.org/556049 | 17:31 |
| evrardjp | logan-: I know that part, but at least it won't break things. | 17:31 |
| logan- | yeah | 17:31 |
| evrardjp | that's all I wanted to test on this patch :) | 17:31 |
| logan- | gotcha | 17:31 |
| *** savvas has joined #openstack-ansible | 17:31 | |
| evrardjp | first no regressions, then hoping it would fix things until we have real multi nodes in gates :) | 17:31 |
| evrardjp | logan-: what's the reason for the sudden docker love? | 17:32 |
| evrardjp | I am not against, I am just curious about the use case | 17:32 |
| logan- | idk just felt like adding it last night because it is a very small change. didnt realize getting testing working was going to be 1/2 day | 17:32 |
| evrardjp | :D | 17:33 |
| logan- | it is handy for me to be able to connect to docker remotely too | 17:33 |
| *** radeks_ has quit IRC | 17:33 | |
| logan- | the docker connection plugin that exists is like the upstream lxc plugin, it can only work on localhost containers | 17:33 |
| evrardjp | doesn't ansible ship already something for docker connections? | 17:33 |
| logan- | yep ^ | 17:33 |
| evrardjp | ok | 17:33 |
| *** savvas has quit IRC | 17:33 | |
| logan- | i think we should try to upstream the connection plugin once it is uncoupled from the strategy plugin in 2.5 | 17:34 |
| *** savvas has joined #openstack-ansible | 17:34 | |
| evrardjp | I agree | 17:38 |
| logan- | its an excellent plugin but requires a lot of upkeep. if we can get wider usage on it, it might help with that. | 17:39 |
| evrardjp | agreed once more | 17:39 |
| *** savvas has quit IRC | 17:39 | |
| *** savvas has joined #openstack-ansible | 17:40 | |
| evrardjp | ansible will strip a lot of things from the core, but I think it would be good to have this included into some kind of agreement with the cores of ansible | 17:40 |
| evrardjp | mutual agreement of shared ownership | 17:41 |
| *** savvas has quit IRC | 17:41 | |
| evrardjp | (because we won't be able to just dump it I think) | 17:41 |
| logan- | yeah they have some kind of thing implemented for community maintained modules that allows external maintenance | 17:41 |
| *** savvas has joined #openstack-ansible | 17:41 | |
| evrardjp | I meant ofc simply rely on them to do the right thing | 17:41 |
| logan- | so i wouldn't expect it'll replace ssh.py but you could have a community module ssh_container.py maybe, where OSA has the ability to merge PRs via ansibot | 17:42 |
| evrardjp | yeah | 17:42 |
| evrardjp | I am fine with pushing that more | 17:44 |
| evrardjp | I will discuss that next week | 17:44 |
| evrardjp | thanks for the reminder | 17:44 |
| logan- | cool. no prob. | 17:44 |
| evrardjp | in the meantime, thanks for the patches on docker support :) | 17:45 |
| evrardjp | and for improvement of the testing! | 17:45 |
| logan- | it is easy thanks to cloudnull's nsenter stuff | 17:45 |
| evrardjp | I have yet to check what's the real difference with a nsenter -a -t and the lxc-attach | 17:46 |
| evrardjp | I expect the env var cleanup doesn't happen | 17:46 |
| evrardjp | but I will read the code | 17:46 |
| *** savvas has quit IRC | 17:46 | |
| logan- | yea thats a good question. ansible runs in a /bin/sh wrapper so id think we have a clear env there | 17:47 |
| logan- | in a /bin/sh wrapper inside the container* | 17:48 |
| *** savvas has joined #openstack-ansible | 17:49 | |
| *** savvas has quit IRC | 17:51 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-plugins master: [WIP] Docker support in ssh connection plugin https://review.openstack.org/555987 | 17:51 |
| *** gkadam has joined #openstack-ansible | 17:52 | |
| *** savvas has joined #openstack-ansible | 17:52 | |
| *** masuberu has quit IRC | 17:54 | |
| *** gkadam has quit IRC | 17:55 | |
| *** savvas has quit IRC | 17:59 | |
| *** savvas has joined #openstack-ansible | 18:01 | |
| *** savvas has quit IRC | 18:03 | |
| *** savvas has joined #openstack-ansible | 18:13 | |
| *** savvas has quit IRC | 18:13 | |
| *** savvas has joined #openstack-ansible | 18:15 | |
| d34dh0r53 | https://review.openstack.org/#/c/556046 looks good to me, are we good to merge? | 18:15 |
| logan- | seems fine to me, can always iterate from here if the problem persists | 18:16 |
| d34dh0r53 | cool | 18:16 |
| *** savvas has quit IRC | 18:17 | |
| *** savvas has joined #openstack-ansible | 18:25 | |
| *** savvas has quit IRC | 18:29 | |
| *** savvas has joined #openstack-ansible | 18:29 | |
| *** haad1 has joined #openstack-ansible | 18:37 | |
| *** savvas has quit IRC | 18:39 | |
| *** savvas has joined #openstack-ansible | 18:44 | |
| *** savvas has quit IRC | 18:49 | |
| *** savvas has joined #openstack-ansible | 18:49 | |
| *** epalper has joined #openstack-ansible | 18:49 | |
| *** savvas has quit IRC | 18:53 | |
| *** savvas has joined #openstack-ansible | 18:59 | |
| *** savvas has quit IRC | 19:04 | |
| *** savvas has joined #openstack-ansible | 19:04 | |
| xdfil | I'm trying to migrate a cinder volume backed VM (iscsi) from a local disk compute host, to a rbd compute host, and getting incompatability response... | 19:06 |
| xdfil | http://libvirt.org/html/libvirt-libvirt-host.html#virCPUCompareResult | 19:06 |
| xdfil | It shouldn't really matter though right? | 19:06 |
| *** epalper has quit IRC | 19:06 | |
| xdfil | the VM has no dependency on the compute hosts instance storage... or does it? | 19:07 |
| *** savvas has quit IRC | 19:10 | |
| logan- | xdfil: probably unrelated to storage | 19:12 |
| xdfil | its the only difference though | 19:13 |
| logan- | are you trying to live migrate a long running instance where the src/dest hosts have different versions of libvirt packages | 19:13 |
| logan- | libvirt/qemu packages* | 19:13 |
| logan- | or rather, the instance was started when older packages were installed | 19:13 |
| xdfil | nope these were all update in the last 24 hours | 19:13 |
| xdfil | and i can migrate rbd backed between rbd backed hosts | 19:14 |
| xdfil | and non-rbd backed between all the other hosts | 19:14 |
| logan- | usually ill see this when the machine type has changed | 19:14 |
| xdfil | just cant do non-rbd backed VM >> rbd host | 19:14 |
| logan- | can you hard reboot it and then migrate | 19:14 |
| xdfil | ya, hard reboot? | 19:15 |
| logan- | to force regen of the instance xml | 19:15 |
| xdfil | same issue | 19:18 |
| *** savvas has joined #openstack-ansible | 19:18 | |
| logan- | odd | 19:19 |
| xdfil | usually when its a libvirt version difference the error will be about a certain CPU feature being missing | 19:20 |
| xdfil | this time the error is InvalidCPUInfo: Unacceptable CPU info: CPU doesn't have compatibility. 0 | 19:21 |
| xdfil | its a different message | 19:21 |
| logan- | identical /proc/cpuinfo in src/dest? | 19:22 |
| xdfil | yup exact same | 19:22 |
| logan- | :/ | 19:23 |
| *** savvas has quit IRC | 19:23 | |
| xdfil | eh, its a lower priority issue for me right now | 19:23 |
| xdfil | I have enough of both types of computes. I just wont change all of them to rbd just yet | 19:24 |
| *** savvas has joined #openstack-ansible | 19:25 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: Execute cinder_backends on the last API server https://review.openstack.org/556046 | 19:29 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-os_cinder stable/queens: Execute cinder_backends on the last API server https://review.openstack.org/556052 | 19:30 |
| logan- | does that need a backport to pike also ^ | 19:31 |
| *** savvas has quit IRC | 19:33 | |
| *** savvas has joined #openstack-ansible | 19:38 | |
| *** savvas has quit IRC | 19:48 | |
| xdfil | "flavor" is a terrible word | 19:51 |
| *** savvas has joined #openstack-ansible | 19:55 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Adding mnaio qemu file backends https://review.openstack.org/549097 | 19:56 |
| *** savvas has quit IRC | 20:08 | |
| *** savvas has joined #openstack-ansible | 20:08 | |
| *** haad1 has quit IRC | 20:08 | |
| *** savvas has quit IRC | 20:17 | |
| *** savvas has joined #openstack-ansible | 20:17 | |
| *** savvas has quit IRC | 20:21 | |
| *** hamza21 has quit IRC | 20:23 | |
| *** savvas has joined #openstack-ansible | 20:38 | |
| *** savvas has quit IRC | 20:43 | |
| *** savvas has joined #openstack-ansible | 20:44 | |
| *** savvas has quit IRC | 20:50 | |
| *** savvas has joined #openstack-ansible | 20:50 | |
| *** savvas has quit IRC | 20:54 | |
| *** savvas has joined #openstack-ansible | 20:56 | |
| *** savvas has quit IRC | 20:59 | |
| *** savvas has joined #openstack-ansible | 21:01 | |
| *** savvas has quit IRC | 21:03 | |
| *** savvas has joined #openstack-ansible | 21:08 | |
| *** savvas has quit IRC | 21:11 | |
| *** savvas has joined #openstack-ansible | 21:14 | |
| *** savvas has quit IRC | 21:15 | |
| *** savvas has joined #openstack-ansible | 21:15 | |
| xdfil | odyssey4me: I wrote comments on https://review.openstack.org/#/c/555981/ but I'm not sure if i'm saving them right heh | 21:17 |
| *** savvas has quit IRC | 21:21 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/pike: Execute cinder_backends on the last API server https://review.openstack.org/556072 | 21:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/pike: Execute cinder_backends on the last API server https://review.openstack.org/556072 | 21:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/ocata: Execute cinder_backends on the last API server https://review.openstack.org/556073 | 21:24 |
| *** savvas has joined #openstack-ansible | 21:24 | |
| *** sawblade6 has joined #openstack-ansible | 21:26 | |
| *** savvas has quit IRC | 21:30 | |
| odyssey4me | xdfil if you've added comments, you need to submit a 'reply' for them to actually be submitted | 21:30 |
| xdfil | Ahh yes! there we go | 21:30 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_nova master: Rebuild venv if missing https://review.openstack.org/556003 | 21:35 |
| *** savvas has joined #openstack-ansible | 21:39 | |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_glance master: Rebuild venv if missing https://review.openstack.org/556005 | 21:40 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_neutron master: Rebuild venv if missing https://review.openstack.org/556013 | 21:41 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_cinder master: Rebuild venv if missing https://review.openstack.org/555999 | 21:42 |
| *** savvas has quit IRC | 21:44 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Add OpenStack CI and docs/releasenotes scaffolding https://review.openstack.org/556074 | 21:46 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Add OpenStack CI and docs/releasenotes scaffolding https://review.openstack.org/556074 | 21:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Add OpenStack CI and docs/releasenotes scaffolding https://review.openstack.org/556074 | 21:51 |
| *** savvas has joined #openstack-ansible | 21:54 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Add OpenStack CI and docs/releasenotes scaffolding https://review.openstack.org/556074 | 22:03 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Add OpenStack CI and docs/releasenotes scaffolding https://review.openstack.org/556074 | 22:03 |
| *** savvas has quit IRC | 22:05 | |
| xdfil | I'm playing with Gerrit, long overdue.... | 22:05 |
| xdfil | look what I made though! | 22:05 |
| xdfil | https://review.openstack.org/#/q/projects:openstack%252Fopenstack-ansible | 22:05 |
| xdfil | a URL that shows me all the OSA (and roles) changes... thats crazy | 22:06 |
| odyssey4me | xdfil ;) https://review.openstack.org/#/q/project:%22%255E(openstack%257Copenstack-infra)/(monitorstack%257Copenstack-ansible%257Cansible-%257Cdevstack-gate).*%22+status:open | 22:06 |
| odyssey4me | well, here's a shorter one without the devstack thing: https://review.openstack.org/#/q/project:%22%255Eopenstack/(monitorstack%257Copenstack-ansible%257Cansible-).*%22+status:open | 22:07 |
| xdfil | Yeah, this is cool heh | 22:07 |
| xdfil | I already like it better than jira lol | 22:08 |
| odyssey4me | haha, it's purpose specific - whereas jira is a bit more generalist | 22:08 |
| *** prometheanfire has quit IRC | 22:10 | |
| xdfil | yeah issue statuses in Jira have bugged me when people have their own opinion of what the status means | 22:13 |
| *** savvas has joined #openstack-ansible | 22:14 | |
| *** savvas has quit IRC | 22:16 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Add OpenStack CI and docs/releasenotes scaffolding https://review.openstack.org/556074 | 22:16 |
| *** savvas has joined #openstack-ansible | 22:16 | |
| *** sawblade_ has joined #openstack-ansible | 22:24 | |
| xdfil | odyssey4me: So this is an example of the new role structure ? https://review.openstack.org/#/projects/openstack/ansible-role-python_venv_build,dashboards/default | 22:27 |
| *** sawblade6 has quit IRC | 22:27 | |
| xdfil | is there something I can read that has an overview of the structure in general ? | 22:28 |
| *** sawblade6 has joined #openstack-ansible | 22:30 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Add OpenStack CI and docs/releasenotes scaffolding https://review.openstack.org/556074 | 22:33 |
| *** sawblade_ has quit IRC | 22:33 | |
| *** savvas has quit IRC | 22:34 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Add OpenStack CI and docs/releasenotes scaffolding https://review.openstack.org/556074 | 22:34 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Add CI/docs/releasenotes scaffolding and basic test https://review.openstack.org/556074 | 22:35 |
| odyssey4me | logan- there's an improvement suggestion in https://review.openstack.org/555981 which simplifies how that's all done | 22:38 |
| *** prometheanfire has joined #openstack-ansible | 22:38 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Add CI/docs/releasenotes scaffolding and basic test https://review.openstack.org/556074 | 22:42 |
| xdfil | odyssey4me: on https://review.openstack.org/555981 doesn't your suggestion conflict with the non-prebuilt venvs? | 22:50 |
| odyssey4me | xdfil whether using developer mode or not - that folder must exist, so there's no need for the folder creation task to have a conditional | 22:50 |
| xdfil | agree, but later on we are also going to unarchive the venv based on if we made the directory | 22:51 |
| odyssey4me | but that's only one of the conditionals | 22:52 |
| xdfil | oh wait is there another conditional | 22:52 |
| xdfil | dang ya | 22:52 |
| xdfil | you right | 22:52 |
| xdfil | See thats what i said to logan- last night lol | 22:52 |
| odyssey4me | yep, it's either if the extract happened OR the directory got created | 22:52 |
| odyssey4me | if both happened, it still activates the task | 22:52 |
| xdfil | yeah when I saw that var I thought to myself that when it was made the author probably had that intent and then forgot | 22:53 |
| xdfil | I was really stuck on that issue lol | 22:54 |
| xdfil | I didn't notice it was the existence of the cached image that was making it skip it was driving me nuts | 22:55 |
| odyssey4me | yeah, it's a good bug to catch that one - a simple oversight, really | 22:58 |
| odyssey4me | with thanks to your report, someone else won't find themselves in the same predicament | 22:58 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_glance master: Rebuild venv if missingIf a deployer removes a venv from /openstack/venvs and runs the role in an attempt to re-deploy the venv from the same tag, the role will fail. This is because the venv archive in /var/cache is still present with the same checksum, https://review.openstack.org/556091 | 23:09 |
| xdfil | damin i did it wrong | 23:10 |
| xdfil | I was trying to update my previous change not submit a new one :) | 23:11 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_glance master: Rebuild venv if missing https://review.openstack.org/556091 | 23:12 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_glance master: Rebuild venv if missing https://review.openstack.org/556091 | 23:14 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_keystone master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556092 | 23:17 |
| odyssey4me | xdfil you need to 'git commit --amend' to ensure that they're all in one commit | 23:17 |
| xdfil | yeah I was being lazy and trying to do it with VScode | 23:18 |
| odyssey4me | patch storm coming in - apologies | 23:18 |
| odyssey4me | you should be able to amend a commit with VScode | 23:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_almanach master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556093 | 23:18 |
| xdfil | you're going to wake the baby! | 23:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_aodh master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556094 | 23:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_barbican master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556095 | 23:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ceilometer master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556096 | 23:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556097 | 23:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cloudkitty master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556098 | 23:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_congress master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556099 | 23:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_designate master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556100 | 23:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_gnocchi master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556101 | 23:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_heat master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556102 | 23:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_horizon master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556103 | 23:20 |
| evrardjp | thanks for the spam odyssey4me ! | 23:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ironic master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556104 | 23:20 |
| evrardjp | it woke me up! :p | 23:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_magnum master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556105 | 23:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_monasca master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556106 | 23:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556107 | 23:20 |
| evrardjp | thanks for the patches :) | 23:20 |
| logan- | thanks @ the notes xdfil odyssey4me will fix that up | 23:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556108 | 23:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_octavia master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556109 | 23:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_rally master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556110 | 23:21 |
| logan- | odyssey4me just to understand better if you have a sec I'd like to know what broke between pike and queens here https://review.openstack.org/#/c/556072/ | 23:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_sahara master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556111 | 23:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_searchlight master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556112 | 23:21 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_cinder master: Rebuild venv if missing https://review.openstack.org/555999 | 23:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556113 | 23:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tacker master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556114 | 23:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556115 | 23:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_trove master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556116 | 23:22 |
| *** dave-mcc_ has quit IRC | 23:22 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_watcher master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556117 | 23:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_zaqar master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556118 | 23:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: ansible-role-requirements: Add common python_venv_build role https://review.openstack.org/556119 | 23:22 |
| odyssey4me | logan- pike has a different service setup order to queens, thanks to a patch from jmccrory - I'm not sure if there was intent to backport it to pike and older | 23:23 |
| odyssey4me | master/queens uses the cinder-api group here: https://github.com/openstack/openstack-ansible-os_cinder/blob/master/tasks/main.yml#L99 | 23:24 |
| odyssey4me | pike and older use the cinder_volume group: https://github.com/openstack/openstack-ansible-os_cinder/blob/stable/pike/tasks/main.yml#L99 | 23:24 |
| odyssey4me | master/queens orders the play with scheduler, volume, backup, api: https://github.com/openstack/openstack-ansible/blob/master/playbooks/os-cinder-install.yml#L60-L88 | 23:25 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_nova master: Rebuild venv if missing https://review.openstack.org/556003 | 23:25 |
| odyssey4me | pike and older does api, scheduler, volume, backup: https://github.com/openstack/openstack-ansible/blob/stable/pike/playbooks/os-cinder-install.yml#L61-L89 | 23:25 |
| odyssey4me | logan- does all that make sense, and highlight why pike's working ok - but queens is not? | 23:27 |
| odyssey4me | jmccrory so I guess something we should follow up on is whether https://review.openstack.org/528140 should be ported back to pike/ocata/newton? | 23:28 |
| odyssey4me | it doesn't seem necessary - perhaps it's more optimal, but perhaps the resulting pain isn't really worth it | 23:28 |
| openstackgerrit | Filippo DiNoto proposed openstack/openstack-ansible-os_neutron master: Rebuild venv if missing https://review.openstack.org/556013 | 23:29 |
| logan- | yeah that makes sense | 23:29 |
| logan- | why it broke makes sense i mean | 23:29 |
| logan- | but it seems like running cinder_backends scoped to cinder_volume makes more sense | 23:30 |
| logan- | because that is where the backends are configured | 23:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Add common python_venv_build role https://review.openstack.org/556119 | 23:30 |
| odyssey4me | logan- yep, and it was previously - until jmccrory submitted https://github.com/openstack/openstack-ansible-os_cinder/commit/6c976e5077b03066f26f8b9d65c9071828419d9f#diff-2444ad0870f91f17ca6c2a5e96b26823 | 23:31 |
| logan- | yea | 23:31 |
| odyssey4me | given the ordering change, it makes sense | 23:32 |
| logan- | ah | 23:32 |
| logan- | yeah since api is deployed last now | 23:32 |
| odyssey4me | yup, it's an odd ordering - but apparently that's what cinder docs suggest as ordering, so meh | 23:33 |
| odyssey4me | I guess it makes sense to implement back-end services, then the API which contacts them. | 23:33 |
| openstackgerrit | Merged openstack/openstack-ansible master: Disable ceph-ansible NTP installation https://review.openstack.org/549638 | 23:34 |
| openstackgerrit | Merged openstack/openstack-ansible master: [Docs] Include OpenStack-Ansible Manifesto https://review.openstack.org/555903 | 23:34 |
| logan- | yeah ¯\_(ツ)_/¯... thanks for explaining why it broke though | 23:35 |
| odyssey4me | evrardjp sorry for waking you up! | 23:35 |
| odyssey4me | logan- oh, it's my pleasure | 23:35 |
| *** prometheanfire has quit IRC | 23:36 | |
| odyssey4me | apologies for not realising all that sooner - I know that a bunch of people were trying to figure out a way around the problem. | 23:36 |
| odyssey4me | But meh, at least we found a way to solve it. :) | 23:37 |
| *** prometheanfire has joined #openstack-ansible | 23:37 | |
| odyssey4me | xdfil could you do one more thing in those reviews to help make it easier to work through them in one go? | 23:38 |
| xdfil | whatsup | 23:38 |
| odyssey4me | set the 'topic' (or git branch) to the same as the one in https://review.openstack.org/555981 | 23:38 |
| odyssey4me | that way it's really easy in gerrit to find all the same patches across the roles and work through them in one go :) | 23:38 |
| odyssey4me | right now they're likely to get lost in the woods of all other patches going on | 23:39 |
| odyssey4me | xdfil make sense? | 23:41 |
| xdfil | odyssey4me: yes, done (i think) | 23:41 |
| odyssey4me | thanks | 23:42 |
| odyssey4me | let's get one of those merged to finalise the pattern - reviewing now | 23:42 |
| xdfil | odyssey4me: was the topic there on the logans original submit or did you add that or? | 23:42 |
| xdfil | or in other words, should I be setting topics all the time lol | 23:43 |
| odyssey4me | xdfil the topic comes from the git branch you use, and typically in a git workflow you would checkout master, update it, then create the new branch which describes what you're going to do briefly | 23:44 |
| odyssey4me | that becomes your 'bug' or 'feature' branch | 23:44 |
| odyssey4me | gerrit takes the branch name and makes it the 'topic' | 23:45 |
| xdfil | so the topic is just make beleive | 23:45 |
| odyssey4me | logan- added that topic for his original patch | 23:45 |
| xdfil | I mean | 23:45 |
| odyssey4me | a topic's a really good way of relating patches | 23:45 |
| odyssey4me | so, for example, if you want to see all the patches that's been done in the python build simplificattion work, you can just use the topic: https://review.openstack.org/#/q/topic:bp/python-build-install-simplification+(status:open+OR+status:merged) | 23:46 |
| xdfil | yeah looks that way | 23:46 |
| odyssey4me | so, generally, if you're doing a one-off patch then a topic isn't necessary | 23:47 |
| xdfil | but within git, the topic is just more characters appended to an existing branch name, to create a new branch | 23:47 |
| odyssey4me | but if you're making a related set of changes across multiple repositories, then the topic is super-handy for reviewers because they can see all the related patches very quickly and easily | 23:48 |
| odyssey4me | nope, in git there is no such thing as a topic - the branch name is translated by gerrit into the topic | 23:48 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-os_keystone master: Rebuild venv if missing https://review.openstack.org/555981 | 23:48 |
| xdfil | but/and since we're using git-review we don't have to make new branches in git because gerrit | 23:49 |
| xdfil | ok i think you made me a little smarter today | 23:50 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-os_keystone master: Rebuild venv if missing https://review.openstack.org/555981 | 23:50 |
| odyssey4me | logan- tbh I think I like the approach done by xdfil in https://review.openstack.org/#/c/555999/3/tasks/cinder_install.yml - it's just ever so slightly easier to follow | 23:52 |
| odyssey4me | hmm, the use of block works too | 23:53 |
| logan- | im going to move the block above the install task | 23:53 |
| xdfil | blocks are just like nesting if statements ? | 23:53 |
| logan- | so it is less broken up | 23:54 |
| odyssey4me | logan- yeah, makes sense | 23:54 |
| xdfil | I've never used a block before hehe | 23:54 |
| odyssey4me | xdfil sort of http://docs.ansible.com/ansible/latest/playbooks_blocks.html | 23:54 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-os_keystone master: Rebuild venv if missing https://review.openstack.org/555981 | 23:54 |
| logan- | you can do other stuff with blocks but yeah one thing is applying the same condition to multiple tasks | 23:55 |
| logan- | there hows that look ^ | 23:55 |
| odyssey4me | logan- not sure if we want to apply this pattern generally, but I find https://github.com/openstack/ansible-role-python_venv_build/blob/master/tasks/python_venv_build.yml#L65-L67 a way to more easily get to grips with a block/conditional | 23:56 |
| odyssey4me | I often find that the conditional gets lost when using it for a block. | 23:56 |
| logan- | so put it at the beginning | 23:56 |
| logan- | i like that | 23:56 |
| odyssey4me | unfortunately that doesn't work with ansible 2.1, because it doesn't support naming blocks - but meh, that's newton :p | 23:57 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-os_keystone master: Rebuild venv if missing https://review.openstack.org/555981 | 23:57 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!