| *** ansmith has joined #openstack-ansible | 00:07 | |
| *** rodolof has joined #openstack-ansible | 00:09 | |
| *** gyee has quit IRC | 00:17 | |
| *** weezS has quit IRC | 00:21 | |
| *** cshen has quit IRC | 00:27 | |
| *** spatel has joined #openstack-ansible | 00:29 | |
| *** spatel has quit IRC | 00:34 | |
| *** tosky has quit IRC | 01:19 | |
| *** cshen has joined #openstack-ansible | 01:21 | |
| *** cshen has quit IRC | 01:25 | |
| *** markvoelker has quit IRC | 01:33 | |
| jamesdenton | redkrieg still around? | 01:48 |
|---|---|---|
| *** mma has joined #openstack-ansible | 01:55 | |
| *** mma has quit IRC | 01:59 | |
| *** spatel has joined #openstack-ansible | 02:22 | |
| redkrieg | jamesdenton: hey, just finished having dinner | 02:34 |
| jamesdenton | no worries. | 02:34 |
| jamesdenton | let me wrap up what i'm doing and i will try and help you out | 02:35 |
| redkrieg | thanks, I'll be around | 02:48 |
| jamesdenton | alright, enough of that | 02:48 |
| redkrieg | oh, hey | 02:48 |
| jamesdenton | So, single controller, single compute | 02:49 |
| jamesdenton | the openvswitch_agent.ini you provided - was that from the compute node? controller? | 02:50 |
| jamesdenton | and, what version of OSA is this? Queens? Rocky? | 02:50 |
| redkrieg | the openvswitch_agent was from the compute node, this is rocky | 02:52 |
| jamesdenton | ok. so one issue i saw was that the [ovs] block was missing some needed information. notably, the bridge_mappings config | 02:53 |
| jamesdenton | which ought to be built out | 02:53 |
| redkrieg | interesting, that exists on the controller but not compute | 02:53 |
| jamesdenton | what does it look like on the controller? | 02:54 |
| redkrieg | bridge_mappings = flat:br-provider | 02:54 |
| jamesdenton | ok, yep. | 02:54 |
| redkrieg | wow, it launched successfully | 02:55 |
| redkrieg | that was it | 02:55 |
| jamesdenton | the neutron-openvswitch-agent.log you provided: was that from controller or compute? | 02:55 |
| jamesdenton | ok, great! now to figure out why it didn't get set :D | 02:55 |
| redkrieg | compute | 02:56 |
| jamesdenton | when you configured that - did you restart the agent? | 02:56 |
| jamesdenton | just curious if the agent was in a restart loop | 02:56 |
| redkrieg | I did restart it | 02:57 |
| redkrieg | force of habit | 02:57 |
| jamesdenton | ok cool | 02:57 |
| jamesdenton | no problem | 02:57 |
| redkrieg | I've been beating my head against this for at least a day, thank you so much. I still don't actually have connectivity, but I think I can probably troubleshoot that after I figure out why the config isn't getting written correctly | 02:58 |
| jamesdenton | ahh ok. i found it | 02:58 |
| jamesdenton | https://github.com/openstack/openstack-ansible-os_neutron/blob/stable/rocky/templates/plugins/ml2/openvswitch_agent.ini.j2#L10-L12 | 02:58 |
| jamesdenton | so basically, the bridge mapping wasn't added because the compute node is not in the neutron-l3-agent group. meaning that the assumption is you would only need provider networks on a network node, and would use vxlan for computes | 02:59 |
| jamesdenton | i don't really like that. | 02:59 |
| jamesdenton | i think there may be a patch out there to address this | 02:59 |
| jamesdenton | actually, it's a regression of sorts | 03:01 |
| redkrieg | ahh yeah, I want my guest traffic going straight out of the computes for now | 03:01 |
| jamesdenton | rightly so | 03:01 |
| jamesdenton | https://bugs.launchpad.net/openstack-ansible/+bug/1807074 | 03:05 |
| openstack | Launchpad bug 1807074 in openstack-ansible "Neutron OVS bridge_mappings not built on computes" [Undecided,New] | 03:05 |
| jamesdenton | Thanks for the heads up! | 03:06 |
| redkrieg | Thank you for the quick diagnosis! Is it safe to add my comp1 node to neutron-l3-agent? and would the best way to do that be putting it in a neutron-l3-agent_hosts section in openstack_user_config.yml? | 03:09 |
| jamesdenton | hmm, good question. you may be able to get away with changing the neutron_plugin_type from ml2.ovs to ml2.ovs.dvr | 03:11 |
| redkrieg | the last time I looked at that plugin was a couple years back, I assume it's a little further along? | 03:11 |
| jamesdenton | we have folks using it in production now, so there's some confidence there. But you don't have to use it - we're just using it to manipulate the inventory here | 03:13 |
| redkrieg | no, I liked the promises it had before, I just had too many conflicts getting it working with my salt-based setup. I'm sure it's matured much since then | 03:14 |
| jamesdenton | gotcha. yeah, on the surface it works well. it's complicated at the OVS layer, but it's been baking since Juno so I hope it's improved :D | 03:15 |
| redkrieg | hopefully it likes my network config, I'm running playbooks and then heading out for the evening so I won't find out until tomorrow. thanks again for all the help! | 03:16 |
| jamesdenton | good deal. i'll be around, so if you still have issues feel free to ask | 03:16 |
| redkrieg | will do | 03:17 |
| *** cshen has joined #openstack-ansible | 03:21 | |
| *** rodolof has quit IRC | 03:23 | |
| *** cshen has quit IRC | 03:26 | |
| *** hwoarang has quit IRC | 03:47 | |
| *** rodolof has joined #openstack-ansible | 03:49 | |
| *** hwoarang has joined #openstack-ansible | 03:50 | |
| *** rodolof has quit IRC | 04:23 | |
| *** cshen has joined #openstack-ansible | 05:22 | |
| *** cshen has quit IRC | 05:26 | |
| *** mma has joined #openstack-ansible | 05:30 | |
| *** ahosam has joined #openstack-ansible | 05:32 | |
| *** spatel has quit IRC | 05:46 | |
| *** aedc has joined #openstack-ansible | 06:35 | |
| *** mma has quit IRC | 06:41 | |
| *** aedc has quit IRC | 06:48 | |
| *** radeks__ has joined #openstack-ansible | 06:49 | |
| *** ahosam has quit IRC | 07:00 | |
| *** markvoelker has joined #openstack-ansible | 07:00 | |
| *** markvoelker has quit IRC | 07:05 | |
| *** radeks_ has joined #openstack-ansible | 07:22 | |
| *** vnogin has joined #openstack-ansible | 07:24 | |
| *** ahosam has joined #openstack-ansible | 07:24 | |
| *** radeks__ has quit IRC | 07:25 | |
| openstackgerrit | jacky06 proposed openstack/openstack-ansible master: Change openstack-dev to openstack-discuss https://review.openstack.org/622646 | 07:33 |
| openstackgerrit | jacky06 proposed openstack/openstack-ansible master: Change openstack-dev to openstack-discuss https://review.openstack.org/622646 | 07:33 |
| *** aedc has joined #openstack-ansible | 07:40 | |
| *** mma has joined #openstack-ansible | 07:43 | |
| *** radeks_ has quit IRC | 07:50 | |
| *** hamzaachi has joined #openstack-ansible | 07:55 | |
| *** pcaruana has joined #openstack-ansible | 07:58 | |
| *** pcaruana is now known as muttley | 07:58 | |
| *** shardy has joined #openstack-ansible | 08:02 | |
| *** shardy has quit IRC | 08:05 | |
| *** faizy_ has joined #openstack-ansible | 08:07 | |
| *** faizy98 has quit IRC | 08:09 | |
| *** gkadam has joined #openstack-ansible | 08:16 | |
| *** faizy_ has quit IRC | 08:23 | |
| *** cshen has joined #openstack-ansible | 08:26 | |
| *** shardy has joined #openstack-ansible | 08:28 | |
| *** rgogunskiy has joined #openstack-ansible | 08:30 | |
| *** aedc has quit IRC | 08:32 | |
| *** dodo_o has quit IRC | 08:45 | |
| *** aedc has joined #openstack-ansible | 08:52 | |
| *** vnogin has quit IRC | 08:54 | |
| *** ahosam has quit IRC | 08:54 | |
| *** rgogunskiy has quit IRC | 08:57 | |
| *** tosky has joined #openstack-ansible | 09:00 | |
| *** ahosam has joined #openstack-ansible | 09:01 | |
| *** markvoelker has joined #openstack-ansible | 09:01 | |
| *** rgogunskiy has joined #openstack-ansible | 09:04 | |
| *** rgogunskiy has quit IRC | 09:08 | |
| chandan_kumar | odyssey4me: Hello | 09:08 |
| chandan_kumar | odyssey4me: https://review.openstack.org/622999 please have a look at this one, I am not sure why centos job timed out | 09:09 |
| chandan_kumar | odyssey4me: http://logs.openstack.org/99/622999/1/check/openstack-ansible-functional-centos-7/52422b7/job-output.txt.gz#_2018-12-06_06_07_20_064402 | 09:12 |
| *** sm806 has quit IRC | 09:14 | |
| *** sm806 has joined #openstack-ansible | 09:14 | |
| *** DanyC has joined #openstack-ansible | 09:24 | |
| *** ahosam has quit IRC | 09:26 | |
| *** ahosam has joined #openstack-ansible | 09:26 | |
| *** DanyC has quit IRC | 09:29 | |
| *** DanyC has joined #openstack-ansible | 09:30 | |
| *** markvoelker has quit IRC | 09:34 | |
| *** sm806 has quit IRC | 09:43 | |
| *** sm806 has joined #openstack-ansible | 09:43 | |
| jrosser | chandan_kumar: as far as i can see that job has just run extraordinarily slowly throughout - theres no particular bit seems to have got bogged down | 09:45 |
| chandan_kumar | jrosser: let me do a safely recheck | 09:46 |
| jrosser | perhaps something to note is that we are now running centos 7.6 images there, which wasnt the case yesterday | 09:47 |
| *** mma has quit IRC | 09:48 | |
| *** vnogin has joined #openstack-ansible | 09:49 | |
| *** mma has joined #openstack-ansible | 10:00 | |
| *** electrofelix has joined #openstack-ansible | 10:04 | |
| *** ahosam has quit IRC | 10:15 | |
| *** markvoelker has joined #openstack-ansible | 10:31 | |
| openstackgerrit | chen jiao proposed openstack/openstack-ansible-os_swift master: spelling error https://review.openstack.org/623181 | 10:35 |
| *** sm806 has quit IRC | 10:59 | |
| *** sm806 has joined #openstack-ansible | 11:00 | |
| *** markvoelker has quit IRC | 11:05 | |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: use tempestconf profile to manage tempestconf cli args https://review.openstack.org/623187 | 11:08 |
| admin0 | \o | 11:10 |
| admin0 | hi all .. missing what step in the deployment will put "# No DNS servers known." in the resolv.conf in the containers ? | 11:10 |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: [WIP] use tempestconf profile to manage tempestconf cli args https://review.openstack.org/623187 | 11:14 |
| *** hamzaachi has quit IRC | 11:20 | |
| *** aedc has quit IRC | 11:49 | |
| *** aedc has joined #openstack-ansible | 11:49 | |
| *** gkadam_ has joined #openstack-ansible | 12:00 | |
| *** aedc has quit IRC | 12:03 | |
| *** gkadam has quit IRC | 12:03 | |
| *** gkadam_ has quit IRC | 12:03 | |
| odyssey4me | morning folks, is centos still causing breakage? | 12:03 |
| *** gkadam has joined #openstack-ansible | 12:05 | |
| *** hamzaachi has joined #openstack-ansible | 12:10 | |
| *** ansmith has quit IRC | 12:11 | |
| *** ahosam has joined #openstack-ansible | 12:27 | |
| *** jferrieu has joined #openstack-ansible | 12:36 | |
| *** udesale has joined #openstack-ansible | 12:41 | |
| canori01 | hey guys, if I already have containers running rbd-backed cinder-volume instances, and add a new node with lvm-backed cinder-volume, will the latter be deployed to metal regardless of the is_metal setting for the container_skel? | 12:52 |
| *** gkadam_ has joined #openstack-ansible | 12:55 | |
| *** cshen has quit IRC | 12:56 | |
| *** vnogin has quit IRC | 12:57 | |
| *** gkadam has quit IRC | 12:58 | |
| guilhermesp | odyssey4me: I didn't see yet but the last update from mnaser was still in progress figuring out the cause | 13:03 |
| guilhermesp | I will take a look within a few minutes | 13:03 |
| odyssey4me | it looked to me like it was still some sort of issue related to the new centos version change and osme sort of mismatch between that and the container | 13:03 |
| guilhermesp | ^ I do remember something related mentioned in openstack-infra | 13:08 |
| guilhermesp | yesterday | 13:08 |
| *** muttley has quit IRC | 13:08 | |
| jrosser | o/ hello odyssey4me | 13:15 |
| odyssey4me | o/ jrosser :) | 13:15 |
| jrosser | this is worrying https://review.openstack.org/#/c/622999/1 | 13:15 |
| *** dave-mccowan has joined #openstack-ansible | 13:16 | |
| odyssey4me | ugh, timeouts - wtf | 13:16 |
| jrosser | i'm slightly concerned about a performance regression with ansible 2.7.4 | 13:16 |
| jrosser | but it could be provider having issues | 13:16 |
| jrosser | need to try to figure out whats going on there | 13:16 |
| odyssey4me | 2.7.4 apparently was mostly packaging related changes... although we jumped from 2.7.2 to 2.7.4 | 13:17 |
| jrosser | .3 was only out for a couple of days, some silly bug made .4 necessary iirc | 13:17 |
| odyssey4me | canori01 you should be able to override the is_metal flag on a per host basis - although I personally have no idea how to | 13:19 |
| odyssey4me | canori01 but the answer is that no, as it stands now it will create it in a container | 13:19 |
| *** dave-mccowan has quit IRC | 13:21 | |
| *** muttley has joined #openstack-ansible | 13:21 | |
| *** muttley has quit IRC | 13:25 | |
| *** muttley has joined #openstack-ansible | 13:26 | |
| guilhermesp | odyssey4me: I will test it locally to see what is going on https://review.openstack.org/#/c/622551/ Any notes on this? | 13:27 |
| odyssey4me | guilhermesp I'd suggest for now, for placement, just include rabbit for keystone's sake - I'll look into it again some time later, but I don't see the harm in having it there. | 13:29 |
| *** muttley has quit IRC | 13:29 | |
| guilhermesp | hum ok odyssey4me I'm going to rollback my PR, seems that we had greenfield before I removed rabbitmq stuff from the role | 13:29 |
| odyssey4me | yeah, it doesn't make sense to delay that work for this issue | 13:30 |
| *** vollman has joined #openstack-ansible | 13:30 | |
| guilhermesp | cool, let me revert and I appreciate any review from you and the team :) | 13:30 |
| *** CeeMac has joined #openstack-ansible | 13:32 | |
| *** pcaruana has joined #openstack-ansible | 13:34 | |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-os_placement master: [WIP] Create base files to install placement https://review.openstack.org/618820 | 13:34 |
| mgariepy | any core want an easy one ? https://review.openstack.org/#/c/622978/ | 13:35 |
| mgariepy | i should have tagged it [doc].. | 13:35 |
| CeeMac | hello channel, I was wondering if anyone was able to answer a couple of (hopefully) quick questions? | 13:36 |
| *** ahosam has quit IRC | 13:37 | |
| mgariepy | CeeMac, the easiest way to know for sure is to ask them :)\ | 13:37 |
| *** ahosam has joined #openstack-ansible | 13:37 | |
| CeeMac | fair play :) | 13:37 |
| CeeMac | so, to set the scene, I've a novice at both openstack and ansible. I've made some good progress deploying AIO, and have started working on a distributed install | 13:38 |
| *** cshen has joined #openstack-ansible | 13:39 | |
| *** pcaruana has quit IRC | 13:39 | |
| CeeMac | I'm a little confused around the haproxy set up and 'external' access networks for neutron, where the host bridge configuration is concerned | 13:39 |
| CeeMac | i'd seen a lot of earlier documentation referencing br-ex, but this seems to be depricated now. | 13:39 |
| jrosser | CeeMac: o/ hello there | 13:40 |
| CeeMac | in the user_variables.yml example file, it references br-flat for the haproxy_keepalived_external_interface | 13:40 |
| jrosser | the example config for openstack ansible assumes that the external network is a single, "untagged" network on br-vlan | 13:40 |
| jrosser | for the neutron external network | 13:40 |
| CeeMac | hi jrosser | 13:41 |
| jrosser | you can do whatever you like with the haproxy external IP, it doesnt have to be in the same subnet at all as the neutron networks, you can do whatever suits your requirements there | 13:41 |
| jrosser | the example configs keep things simple, and combine those ranges | 13:42 |
| CeeMac | that makes sense, say I wanted to keep them on the same subnet though, would they need separate physical nic configurations? | 13:42 |
| CeeMac | or a un-numbered nic, and a bridge with an ip? | 13:43 |
| CeeMac | neutron prefers unnumbered, but haproxy would need a numbered nic with gateway? | 13:44 |
| *** pcaruana has joined #openstack-ansible | 13:44 | |
| jrosser | perhaps it would be worth taking some time to dig through this https://github.com/openstack/openstack-ansible-ops/tree/master/multi-node-aio | 13:44 |
| jrosser | thats a test multnode deploy that runs each node in a VM | 13:44 |
| jrosser | you should be able to find in there how each node has set up | 13:44 |
| jamesdenton | mornin' | 13:45 |
| CeeMac | cool, I'll go digging there. I've run myself around in circles looking through various docs for the different projects etc | 13:45 |
| *** ansmith has joined #openstack-ansible | 13:46 | |
| CeeMac | as an aside, is there a configuration override to inject a static route into keepalived via user_variables.yml? Couldn't see anything in the configuration reference | 13:46 |
| *** pcaruana has quit IRC | 13:47 | |
| *** ahosam has quit IRC | 13:47 | |
| jrosser | CeeMac: see right down the bottom of here https://docs.openstack.org/openstack-ansible/rocky/user/prod/example.html | 13:47 |
| jrosser | thats what you need for defining how/which interface keepalived uses for a multinode deploy | 13:48 |
| jrosser | and so that is where the haproxy external VIP will be | 13:48 |
| CeeMac | yep, saw that, thanks. Maybe I'm over complicating things in my head. I was working on the principle of the interface (ens2 in the example) not having an IP, therefore I'd need to inject a static route for the gateway as part of the keepalived config | 13:49 |
| jamesdenton | CeeMac How many network interfaces are you working with? | 13:50 |
| CeeMac | maybe it would be more straightforward to have a separate interface for haproxy than for neutron external access | 13:50 |
| CeeMac | i can have as many as need be. I'm modelling this in vmware right now, so can have up to 10 | 13:50 |
| CeeMac | but irl there would be 4x10GB and min 4x1GB | 13:50 |
| admin0 | CeeMac, all that haproxy will do is bind the ip to the interface as secondary .. so any ethX or bridge will work out fine | 13:51 |
| CeeMac | working on using bonded pairs | 13:51 |
| admin0 | and they even don't have to be in any subnets or known subnets | 13:51 |
| CeeMac | @admin0 i saw the IP was added as a secondary to the interface, was just trying to work out the path of return traffic if there was a gateway already existing on a different interface/bridge | 13:52 |
| CeeMac | would want return traffic to egress the same interface it ingresses on | 13:53 |
| CeeMac | maybe i'm over-complicating things :s | 13:53 |
| admin0 | for example, i have 4 brides and one say eth10 with 192.168.0.1 .. what i do is give eth10 as my external interface .. and it listens to that .. now how traffic comes inside eth10 and how it leaves is beyond the role of osa | 13:54 |
| CeeMac | sure | 13:54 |
| admin0 | and also, return path does not necessarily have to be the same as incoming oath | 13:54 |
| admin0 | if your (public) can ping and get a reply from that external IP, traffic will work out as well | 13:55 |
| CeeMac | true, providing any upstream firewalls can deal with that. ofc thats out of scope for osa, I appreciate that. | 13:55 |
| jamesdenton | CeeMac here's an example of an infra node on esx, if it helps: http://paste.openstack.org/show/736759/ | 13:56 |
| admin0 | for example, if your laptop can ping this external IP which you want to give to external endpoint, and you get the response back, then it will work out fine .. | 13:56 |
| CeeMac | thanks jamesdenton, I'll take a look | 13:56 |
| jamesdenton | ens224 (not shown) is used for OVS provider bridge in that environment | 13:56 |
| admin0 | CeeMac, in that example, if your clients can ping that 10.50.0.11 when its added to that interface and it works, your api/horizon will work fine as wel | 13:57 |
| *** vnogin has joined #openstack-ansible | 13:57 | |
| CeeMac | thanks guys, i'll take a look at those examples and see if I can decode them and make them work in my environment | 13:57 |
| admin0 | jamesdenton, i am doing osa+ovs like this way: https://www.openstackfaq.com/openstack-ansible-with-openvswitch/ .. do you have a better way or recommendation .. now that i want to use 18.1.0 tag . | 13:58 |
| CeeMac | mainly, I need to be able to put the 'public' network and the external ha proxy interface on an internet routable network, so I'll need a dedicated NIC/bridge for that and keep internal flat/vlan/vxlan traffic on separate nic/bridge | 13:58 |
| admin0 | CeeMac, or you can still use private IP , but SNAT/DNAT that as wel | 13:59 |
| CeeMac | yeah, but that makes it awkward for managing the upstream NAT to public IPs | 14:00 |
| jamesdenton | yes, it does. | 14:00 |
| CeeMac | I'll ultimately be building a multi-tenant setup | 14:00 |
| jamesdenton | if you have the IPs, go for it | 14:00 |
| CeeMac | so wanted to make it as transparent to the project users as possible | 14:00 |
| CeeMac | jamesdenton, you're using the "new" network management there? | 14:01 |
| jamesdenton | admin0 i would just make sure your group_binds are updated to only ovs and no linuxbridge | 14:01 |
| jamesdenton | oh yeah, netplan? | 14:02 |
| CeeMac | thats the badger. | 14:02 |
| jamesdenton | it's not ideal IMO. but it was quick | 14:02 |
| *** rodolof has joined #openstack-ansible | 14:02 | |
| CeeMac | haven't got my head around that yet! | 14:02 |
| jamesdenton | you need ifupdown config? | 14:02 |
| CeeMac | if you have it, please | 14:02 |
| jrosser | fwiw you can still apt install ifupdown and just reuse config you may have from xenial | 14:02 |
| jrosser | ^ i do this | 14:02 |
| CeeMac | jrosser, i just paniced and redeployed with 16.04 instead. my learning curve is already pretty steep jumping in to this project | 14:03 |
| jamesdenton | CeeMac This is a single interface example. If you're using keepalived you should be able to remove the post-up commands. https://gist.github.com/busterswt/01e706086c6bc6108e66e06fc992d1c0 | 14:04 |
| jrosser | no worries - although if you are just starting out then i'd recommend trying to get 18.04 going | 14:04 |
| jamesdenton | Yeah, if you're doing Rocky won't 18.04 be required? | 14:04 |
| CeeMac | yeah, i'll probably revisit that once I bottom out the osa deployment | 14:04 |
| jrosser | rocky is the transition release | 14:05 |
| jamesdenton | oh ok | 14:05 |
| CeeMac | i'd started on queens | 14:05 |
| CeeMac | been doing a looooooot of reading and tinkering over the last couple of months, in between my 'normal' work | 14:05 |
| jrosser | CeeMac: how's your nuke-and-start-again plan? Not being too precious about your install and easy starting again is a huge bonus | 14:06 |
| mgariepy | CeeMac, 18.04 and rocky would be better at this point. if you go 16.04, queens, you will need to upgrade to rocky at some point, and then upgrade the OS. | 14:06 |
| CeeMac | hmm, that wouldn't be ideal | 14:06 |
| CeeMac | jrosser, I'm about ready to nuke-and-redeploy anyway | 14:06 |
| jrosser | excellent :) | 14:06 |
| jamesdenton | CeeMac in that gist, eth0.40 is the management interface of the server. Also using the same interface to apply the 'external' vip. You can use eth0 there if you want, and use an eth1 for the br-vlan bridge. Having a dedicated interface/bond for neutron is recommended. | 14:06 |
| CeeMac | i'd hit a wall with neutron agents being dead | 14:06 |
| *** rodolof has quit IRC | 14:07 | |
| CeeMac | thanks jamesdenton | 14:07 |
| CeeMac | i hadn't picked up on that recommendation, but it makes sense | 14:08 |
| *** rodolof has joined #openstack-ansible | 14:08 | |
| CeeMac | in production, i'd probably run haproxy on dedicated hosts, and neutron on dedicated hosts anyway | 14:08 |
| jamesdenton | i've seen it every way under the sun. | 14:09 |
| CeeMac | to limit the external access, and to make it easier to swap out haproxy for 'physical' LBs at some point | 14:09 |
| jamesdenton | Our haproxy implementation has grown on me | 14:11 |
| jrosser | CeeMac: I do something very similar with a "lots of small hosts" approach, and dedicated bridges/nic/switch and so on for the outside netwroks | 14:11 |
| CeeMac | seems tidier / easier for compliance that way | 14:11 |
| jrosser | but have stuck with the regular haproxy as it seems to work really nicely | 14:11 |
| *** spatel has joined #openstack-ansible | 14:13 | |
| *** spatel has quit IRC | 14:13 | |
| *** spatel has joined #openstack-ansible | 14:13 | |
| *** cshen has quit IRC | 14:17 | |
| *** spatel has quit IRC | 14:17 | |
| CeeMac | as an aside, is there a way to specify the subnet used in the 'public' network that's deployed by osa? | 14:21 |
| *** cshen has joined #openstack-ansible | 14:24 | |
| jrosser | CeeMac: can you be a little more precise? for haproxy/api stuff or for the neutron external network? | 14:27 |
| CeeMac | sorry, the neutron external network | 14:27 |
| CeeMac | its called public | 14:27 |
| jrosser | i think that in that case the OSA specific part is concerned with setting up the information for neutron about which physical interface that is on | 14:28 |
| CeeMac | its been given 10.1.13.0/24 in my deploymeny, presumably thats configurable somewhere | 14:29 |
| CeeMac | as thats not a subnet i have in use | 14:29 |
| jamesdenton | i've not known OSA to configure the network via Neutron API, except maybe for testing | 14:30 |
| jrosser | you've probably run the tempest tests then | 14:30 |
| jrosser | https://github.com/openstack/openstack-ansible-os_tempest/blob/master/defaults/main.yml#L137 | 14:30 |
| CeeMac | oh. i ran the tempest playbook, but didn't run any tests yet | 14:30 |
| jrosser | it will have created the resources required for the tests including an external network with the defaults from that file, unless you override them otherwise | 14:31 |
| CeeMac | got you | 14:32 |
| jrosser | so thats a good example of the division of the setup - OSA configures neutron telling it about the physical interface mappings, "something else" then has to create the logical networks which is either the tempest role or the admin of the cloud via the api/cli/ansible/...... | 14:32 |
| CeeMac | ok, that makes a lot of sense | 14:33 |
| CeeMac | it was the osa tempest playbook i ran though, so presumably there is a config override value somewhere? | 14:33 |
| *** vakuznet has joined #openstack-ansible | 14:34 | |
| jrosser | the external interfaces to all those roles are the defaults/main.yml files | 14:34 |
| odyssey4me | CeeMac that defaults file contains everything possible to override, and all overrides go into user_variables.yml or your user space host/group vars | 14:35 |
| jrosser | so you can look through those for the things you can override, and use your user_...... | 14:35 |
| CeeMac | ah, right. got you. | 14:35 |
| jrosser | ^that :) | 14:35 |
| vakuznet | please review https://review.openstack.org/#/c/621688/ | 14:35 |
| CeeMac | thanks, you've all been really helpful :) | 14:35 |
| CeeMac | i'll go ponder some more and adjust my set up accordingly | 14:35 |
| admin0 | CeeMac, i have around 10 different osa platforms now .. every lb is a different story | 14:36 |
| CeeMac | i guess I'd tried to merge the 2 things into a single entity. Now I know they are distinct I can move forwards hopefully | 14:37 |
| admin0 | CeeMac, also what osa does from the config is maps vlan to br-vlan ( name) .. that is all it does .. after the deployment is done,then you tell neutron .. i have x.x/y ip range .. route it via the vlan interface | 14:37 |
| admin0 | i have br-ext br-dmz br-lan .. so br-ext = public ip, br-dmz goes via firewalls br-lan via ipsec to corporate vpn | 14:38 |
| CeeMac | admin0, is there any reason i can't take the br-vlan 'flat' configuration and set up a seperate br-flat for that? | 14:38 |
| admin0 | CeeMac, there is no hard and fast rule in here .. but even if br-flat is ther ein the config, as long as you don't use it .. having it there in the config does not do anything | 14:39 |
| CeeMac | as long as the openstack_user_config matches with the network config on the host/ | 14:39 |
| admin0 | i have never removed it, but never used it as well | 14:39 |
| CeeMac | so, fire up the basic set up with osa, then use neutron to set up the external network and bind it to the nic/bridge i want? | 14:40 |
| admin0 | yep | 14:40 |
| jrosser | CeeMac: one thing to consider with a flat external network is that you are in a corner there as there can only be one subnet | 14:41 |
| CeeMac | then work out how to reverse-engineer to make that programatically redeployable XD | 14:41 |
| admin0 | CeeMac, why use flat when you can use vlans ? | 14:41 |
| CeeMac | yeah, this is specifically for the 'internet' access | 14:41 |
| jrosser | you would be better off (imho) making it a vlan type and bringing the external network in tagged from your switch | 14:41 |
| admin0 | if you need a router, you can use a virtual one | 14:41 |
| CeeMac | actually | 14:41 |
| CeeMac | yeah | 14:41 |
| CeeMac | i'll be having a seperate switching infrastructure for the public subnets though | 14:42 |
| jrosser | then when the invitable day comes and someone asks you to bring in a second subnet it is trivial | 14:42 |
| admin0 | CeeMac, this is how I usually setup .. https://www.openstackfaq.com/openstack-private-cloud-architecture/ -- either in 1 box or multple .. i use a vyos as the gateway to pxe/dhcp/terminte the vlan and then route from there | 14:42 |
| CeeMac | but you're right, if i do it through vlans, i can allocate different external subnets to different projects if i want | 14:42 |
| jrosser | enable the tag on the switch, api calls to neutrhon and bingo | 14:42 |
| jrosser | *neutron | 14:42 |
| CeeMac | so, ideally I need 2 vlan bridges ideally?, e.g. br-vlan_int and br-vlan_ext | 14:44 |
| admin0 | no need :) its a VLAN :D | 14:45 |
| admin0 | you can from your router/switch route vlan 101 to external and 102 to external | 14:45 |
| CeeMac | but br-vlan is bound to a bond, which has specific nics | 14:45 |
| admin0 | or 103 to vpn | 14:45 |
| admin0 | right .if yuo want separate nics | 14:45 |
| CeeMac | then internal and external vlans will be in separate switching domains | 14:45 |
| CeeMac | yeah | 14:45 |
| CeeMac | i need to maintain separation for compliance | 14:46 |
| CeeMac | makes life easier | 14:46 |
| CeeMac | right now i would do it with dmz vlan and NAT using upstream DC firewall | 14:46 |
| CeeMac | but I'm bringing in a new edge platform so i can deliver public ips direct to VMs or physical devices | 14:47 |
| jrosser | CeeMac: my setup is the same | 14:47 |
| jrosser | note that neutron is actually wanting an *interface* | 14:47 |
| jrosser | not a bridge | 14:47 |
| CeeMac | right | 14:47 |
| CeeMac | does a bond count? | 14:47 |
| admin0 | CeeMac, https://www.openstackfaq.com/openstack-add-direct-attached-dhcp-ip/ -- this is how i have for cpanel vms :) | 14:47 |
| jrosser | which is why you see references to eth12 through the example documentation | 14:47 |
| admin0 | no need to crate network or routers .. direct public IP to the instance | 14:48 |
| CeeMac | oh, thats the reason? I was wondering | 14:48 |
| admin0 | only downsize = if deleted, the ip is lost | 14:48 |
| jrosser | that is a side effect of having the netowrk/controller nodes collapsed and re-using br-vlan for the external (flat) network | 14:48 |
| admin0 | this example provides kind of flat network, but under a vlan | 14:48 |
| jrosser | however, if you have dedicated network nodes, you can just use the external bond directly in the neutron config | 14:48 |
| jrosser | network_mappings: "physnet_rd:bond1" <- there is mine | 14:49 |
| jrosser | network_vlan_ranges: "physnet_rd:1:4000" <- i'm able to specify any vlan to neutron as an external one but the switch config defines those present in the trunk | 14:50 |
| jamesdenton | yeah, ultimately it's up to the agent to figure out how the interface is implemented. linuxbridge might tag bond1 and put it in a brqXXXX bridge. While ovs does expect a bridge name, plus the vm into br-int, and and send traffic out br-ex or br-provider or whatever the physnet_rd mapping goes to | 14:51 |
| admin0 | i normally give 1:4090 in my config .. and deal with networking later :) | 14:51 |
| jrosser | CeeMac: i think that this means if you get stuck just shout :) | 14:52 |
| admin0 | :D | 14:52 |
| admin0 | we talk a lot :) | 14:52 |
| CeeMac | haha, thanks, I'll probably take you up on that offer at some point :D | 14:52 |
| CeeMac | looking at OVS is on my list as well | 14:53 |
| CeeMac | but want to get a working environment up and running with linuxbridge first | 14:53 |
| jamesdenton | good idea. | 14:54 |
| *** francois has joined #openstack-ansible | 14:54 | |
| CeeMac | looping back to 18.04/rocky conversation. Is OSA supporting netplan now? I read some forum posts where people were having problems | 14:54 |
| jamesdenton | I'm not sure OSA cares one way or the other. | 14:54 |
| jamesdenton | i'm not sure happy with it, but i use it | 14:54 |
| jamesdenton | *super | 14:54 |
| CeeMac | ok | 14:54 |
| admin0 | it needs to see the bridges .. does not care how/who created them | 14:55 |
| CeeMac | i'll see if I can make sense of it then | 14:55 |
| CeeMac | i'll probably rebuild my current rc environment as queens/16.04 and rc2 with rocky/18.04, then i can work on upgrade testing later | 14:56 |
| *** aedc has joined #openstack-ansible | 14:57 | |
| CeeMac | ok, thanks again ppl, I'll go look through all of those examples read back through the conversation and see what changes I need to make | 14:57 |
| jamesdenton | good luck! | 14:58 |
| CeeMac | thanks! | 14:59 |
| odyssey4me | CeeMac we use systemd-networkd for everything we setup, because it works on all our supported platforms | 15:03 |
| odyssey4me | But for your bridges and stuff on the host, it's up to you. | 15:04 |
| odyssey4me | the things we setup are the connections from host to container, and inside the container | 15:04 |
| *** mkuf has joined #openstack-ansible | 15:05 | |
| CeeMac | thanks odyssey4me | 15:15 |
| admin0 | jamesdenton, thanks for writing this: http://www.jimmdenton.com/osa-sriov/ | 15:15 |
| *** gkadam_ has quit IRC | 15:19 | |
| jamesdenton | sure! | 15:21 |
| *** hamzaachi has quit IRC | 15:22 | |
| *** jhesketh has quit IRC | 15:34 | |
| *** jhesketh has joined #openstack-ansible | 15:35 | |
| *** strattao has joined #openstack-ansible | 15:41 | |
| mnaser | okay so my progress | 15:45 |
| mnaser | http://logs.openstack.org/94/620994/1/check/openstack-ansible-functional-centos-7/0cb1a96/job-output.txt.gz#_2018-12-06_14_50_05_497015 | 15:45 |
| mnaser | now both host and container are 7.5 | 15:46 |
| mnaser | *7.6 | 15:46 |
| jrosser | i saw some of those | 15:46 |
| mnaser | woo i see placement jobs pass tho | 15:48 |
| mnaser | but no tempest jobs for placements run | 15:48 |
| *** strattao has quit IRC | 15:49 | |
| *** vnogin has quit IRC | 15:51 | |
| *** vnogin has joined #openstack-ansible | 15:51 | |
| *** weezS has joined #openstack-ansible | 15:56 | |
| openstackgerrit | weizj proposed openstack/openstack-ansible-os_searchlight master: Revert "use include_tasks instead of include" https://review.openstack.org/623232 | 15:56 |
| odyssey4me | mnaser so we're all clear for centos now and just need rechecks? or is there something else still broken? | 15:58 |
| mnaser | odyssey4me: i havent seen enough jobs run to know if some of the failures we're seeing are transient or actaul 7.5 failures | 15:58 |
| mnaser | a recheck of an integrated change + repo change might be good exercise | 15:59 |
| *** blinkiz has quit IRC | 15:59 | |
| openstackgerrit | Maxime Guyot proposed openstack/openstack-ansible-ceph_client stable/pike: Remove the dependency on SSH for monitors https://review.openstack.org/623233 | 16:00 |
| openstackgerrit | Maxime Guyot proposed openstack/openstack-ansible-ceph_client stable/rocky: Remove the dependency on SSH for monitors https://review.openstack.org/623234 | 16:00 |
| odyssey4me | mnaser ok, done a few of those - let's see how they go | 16:01 |
| openstackgerrit | Maxime Guyot proposed openstack/openstack-ansible-ceph_client stable/queens: Remove the dependency on SSH for monitors https://review.openstack.org/623235 | 16:01 |
| mnaser | odyssey4me: cool i'll keep following up, sorry for the blockages, trying to fix things the fastest as i can to unblock us | 16:01 |
| openstackgerrit | Maxime Guyot proposed openstack/openstack-ansible-ceph_client stable/ocata: Remove the dependency on SSH for monitors https://review.openstack.org/623236 | 16:04 |
| mnaser | the issues resolved so far are: getting the container images to become 7.6, and now the actual os images in upstream to become 7.6 | 16:05 |
| Miouge | logan-: in Pike I ran into the issue you fixed in 504094 so I figured I would try to backport that into PQR I hope it’s all right | 16:05 |
| odyssey4me | mnaser oh yes, that would be an issue | 16:06 |
| logan- | Miouge: thank you! | 16:06 |
| mnaser | so i chased down centos to get 7.6 containers and now the hosts are 7.6 everywhere with infra (there was a build issue going on) | 16:06 |
| odyssey4me | I started working on just using DIB so that we don't rely on upstream images any more - unfortunately I've been side-tracked, but will try to get that done for Stein at least so this isn't an issue in the future. | 16:06 |
| logan- | Miouge: worked for you on Pike? I expect that patch will work just fine on everything Pike and newer | 16:06 |
| Miouge | logan-: Pipeline is in progress so I don’t know if it worked yet :D | 16:07 |
| *** ThiagoCMC has joined #openstack-ansible | 16:07 | |
| logan- | Miouge: ahh gotcha. ceph_client has no real testing, but 'check experimental' will exercise the role using the full integrated OSA build | 16:08 |
| odyssey4me | logan- unfortunately that experimental check won't exercise the ceph check though :p | 16:08 |
| Miouge | logan-: I meant the pipeline to my staging setup | 16:08 |
| ThiagoCMC | Hey guys, I'm planning to deploy OpenStack Ansible in multiple regions. But my users wants a single Horizon, is it possible? I'm not planning to share Keystone, since we have LDAP and all Keystones can reach that single LDAP (so, LDAP is the shared part). | 16:09 |
| logan- | Miouge: gotcha | 16:09 |
| logan- | odyssey4me: oh really? I thought it ran a ceph job too :/ | 16:09 |
| logan- | dang | 16:09 |
| odyssey4me | logan- well, I think we can make it do so now if we like | 16:09 |
| odyssey4me | lemme push up a patch real quick | 16:09 |
| logan- | we should probably just make that test the gate for ceph_client since the current tests are noop | 16:10 |
| logan- | thank you | 16:10 |
| logan- | at least having experimental will help | 16:10 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_ceilometer master: Add ability to override meters.yaml and event_defenitions.yaml https://review.openstack.org/623239 | 16:13 |
| odyssey4me | logan- that's exactly what I'm about to do | 16:16 |
| odyssey4me | logan- oh, that experimental one does do the integrated ceph deploy - my mistake | 16:20 |
| admin0 | hi logan- , i think you know this: missing what step in the deployment will put "# No DNS servers known." in the resolv.conf in the containers ? | 16:21 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_keystone master: Force force-tlsv12 only https://review.openstack.org/623240 | 16:22 |
| ThiagoCMC | admin0, missing you brother! :-D | 16:22 |
| admin0 | hey ThiagoCMC | 16:22 |
| ThiagoCMC | Hey!! ^_^ | 16:22 |
| admin0 | i was there in CA for 2 weeks .. could not meet you .. you don't reply on skype | 16:22 |
| admin0 | now next time :) | 16:23 |
| ThiagoCMC | Sure, next time! lol | 16:24 |
| ThiagoCMC | Besides a single Horizon for multiple regions, I'm also looking to provide "Ceph as a Service", maybe via Manila and I just found this bug here: https://bugs.launchpad.net/openstack-ansible/+bug/1785386 | 16:26 |
| openstack | Launchpad bug 1785386 in openstack-ansible "Integration of Swift and Manila with Openstack-ansible having Ceph backend?" [Wishlist,Confirmed] | 16:26 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Zuul: Simplify the integrated test playbooks https://review.openstack.org/623250 | 16:36 |
| Miouge | logan-: Mmm something went wrong: http://paste.openstack.org/show/736772/ with the “is success” condition | 16:37 |
| logan- | Miouge: you might have to use '| success' in the older ansible version | 16:37 |
| logan- | i'm not sure when the new syntax became valid | 16:37 |
| logan- | 2.5 maybe? | 16:37 |
| logan- | admin0: doesn't look like anything OSA related places that text: http://codesearch.openstack.org/?q=No%20DNS%20servers%20known&i=nope&files=&repos= | 16:38 |
| admin0 | logan-, :D | 16:38 |
| odyssey4me | logan- admin0 nope, more than likely resolveconf doing it | 16:38 |
| odyssey4me | it can't find DNS servers via DHCP | 16:38 |
| mnaser | odyssey4me: http://logs.openstack.org/04/622304/1/check/openstack-ansible-functional-centos-7/62e4d17/job-output.txt.gz | 16:39 |
| mnaser | job that just posted results 22 minutes ago | 16:39 |
| logan- | yeah perhaps try restarting the lxc dnsmasq on the host admin0 | 16:39 |
| mnaser | it's stable/rocky.. but still | 16:39 |
| admin0 | hmm.. so aio is not foolproof :) | 16:39 |
| Miouge | logan-: OK, i will check that. Thanks | 16:40 |
| admin0 | thanks .. checing | 16:40 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ceph_client master: Test using integrated build https://review.openstack.org/623253 | 16:40 |
| odyssey4me | mnaser ok, so looking promising | 16:41 |
| *** joabdearaujo has joined #openstack-ansible | 16:48 | |
| Miouge | can I edit a cherry picked patch (is success to | succedd) or does it have to be as is? | 17:06 |
| odyssey4me | Miouge you may edit it | 17:07 |
| odyssey4me | Miouge https://docs.openstack.org/openstack-ansible/latest/contributor/contribute.html#backporting | 17:07 |
| *** mma has quit IRC | 17:21 | |
| *** mma has joined #openstack-ansible | 17:21 | |
| *** strattao has joined #openstack-ansible | 17:26 | |
| *** mma has quit IRC | 17:26 | |
| *** cshen has quit IRC | 17:32 | |
| *** jferrieu has quit IRC | 17:36 | |
| *** hamzaachi has joined #openstack-ansible | 17:39 | |
| redkrieg | hey jamesdenton, I tried the dvr plugin we discussed last night but that gave me errors about binding to a dead agent even after reboots and checking services, so I reverted back, disabled the l3 agent that was new, applied playbooks, fixed config, and rebooted. | 17:39 |
| redkrieg | here's my current config if you don't mind giving it a once-over for anything obvious: https://gist.github.com/RedKrieg/d12e3c1d2e0775eaa1f59e89dd999580 | 17:39 |
| redkrieg | the errors from when I had the dvr plugin set up are in the comment at the bottom | 17:40 |
| guilhermesp | hum timeout http://logs.openstack.org/20/618820/48/check/openstack-ansible-functional-centos-7/63ac1b0/job-output.txt.gz would be related to latest centos issues? | 17:41 |
| odyssey4me | guilhermesp maybe, retry - some are getting through | 17:43 |
| openstackgerrit | Ben Hundley proposed openstack/openstack-ansible-os_designate master: Expose default_limit_v2 as variable https://review.openstack.org/623270 | 17:43 |
| *** DanyC has quit IRC | 17:43 | |
| *** hamzaachi has quit IRC | 17:44 | |
| *** hamzaachi has joined #openstack-ansible | 17:44 | |
| *** CeeMac has quit IRC | 17:48 | |
| jrosser | odyssey4me: http://grafana.openstack.org/d/4-pSHcImz/openstack-ansible?orgId=1&from=now%2FM&to=now | 17:54 |
| jrosser | ^ mnaser | 17:54 |
| jrosser | added all the distro build time / fail rates | 17:54 |
| *** gyee has joined #openstack-ansible | 17:55 | |
| mnaser | That’s awesome jrosser | 17:56 |
| mnaser | Let’s see that the numbers add up to be | 17:56 |
| jrosser | i want to add timeout counts too | 17:57 |
| *** vnogin has quit IRC | 17:57 | |
| *** macza has joined #openstack-ansible | 17:57 | |
| *** mma has joined #openstack-ansible | 17:59 | |
| *** mma has quit IRC | 18:00 | |
| *** udesale has quit IRC | 18:04 | |
| *** spatel has joined #openstack-ansible | 18:21 | |
| *** cshen has joined #openstack-ansible | 18:25 | |
| *** electrofelix has quit IRC | 18:26 | |
| *** dave-mccowan has joined #openstack-ansible | 18:30 | |
| spatel | did anyone noticed this bug ? | 18:55 |
| spatel | https://bugs.launchpad.net/horizon/+bug/1807251 | 18:55 |
| openstack | Launchpad bug 1807251 in OpenStack Dashboard (Horizon) "Horizon Overview summary showing wrong numbers " [Undecided,New] | 18:55 |
| *** shardy has quit IRC | 18:59 | |
| *** vollman has quit IRC | 19:04 | |
| *** strattao has quit IRC | 19:04 | |
| *** udesale has joined #openstack-ansible | 19:11 | |
| *** aedc has quit IRC | 19:14 | |
| *** strattao has joined #openstack-ansible | 19:23 | |
| openstackgerrit | Maxime Guyot proposed openstack/openstack-ansible-ceph_client stable/pike: Remove the dependency on SSH for monitors https://review.openstack.org/623233 | 19:28 |
| openstackgerrit | Maxime Guyot proposed openstack/openstack-ansible-ceph_client stable/queens: Remove the dependency on SSH for monitors https://review.openstack.org/623235 | 19:28 |
| *** strattao has quit IRC | 19:30 | |
| *** udesale has quit IRC | 19:31 | |
| *** strattao has joined #openstack-ansible | 20:01 | |
| *** strattao has quit IRC | 20:09 | |
| redkrieg | I was able to get the dvr plugin to start by adding neutron_l2_population: True in user_variables.yml, instances create with that but I still am not getting dhcp relayed to the container's interface :| | 20:18 |
| *** cshen has quit IRC | 20:20 | |
| *** cshen has joined #openstack-ansible | 20:23 | |
| *** mma has joined #openstack-ansible | 20:24 | |
| *** hamzaachi has quit IRC | 20:27 | |
| *** hamzaachi_ has joined #openstack-ansible | 20:28 | |
| spatel | I have submit this bug https://bugs.launchpad.net/openstack-ansible/+bug/1807268 | 20:28 |
| openstack | Launchpad bug 1807268 in openstack-ansible "CentOS rsyslog bug" [Undecided,New] | 20:28 |
| *** mma has quit IRC | 20:28 | |
| *** udesale has joined #openstack-ansible | 20:29 | |
| *** cshen has quit IRC | 20:30 | |
| *** cshen has joined #openstack-ansible | 20:31 | |
| *** DanyC has joined #openstack-ansible | 20:31 | |
| *** DanyC has quit IRC | 20:35 | |
| jamesdenton | hey redkrieg. If i'm looking at your output correctly, you'll want enp4s0f0 in the br-provider bridge and not br-vlan, especially if br-provider is what is listed in that bridge_mappings config in openvswitch_agent.ini | 20:46 |
| *** udesale has quit IRC | 20:46 | |
| openstackgerrit | Frank Kloeker proposed openstack/openstack-ansible-haproxy_server master: Add feature Letsencrypt SSL certification https://review.openstack.org/586774 | 20:59 |
| *** strattao has joined #openstack-ansible | 21:13 | |
| *** strattao has quit IRC | 21:27 | |
| *** CeeMac has joined #openstack-ansible | 21:33 | |
| *** ansmith has quit IRC | 21:41 | |
| openstackgerrit | Frank Kloeker proposed openstack/openstack-ansible-haproxy_server master: Add feature Letsencrypt SSL certification https://review.openstack.org/586774 | 21:47 |
| guilhermesp | at least for placement pr, the centos jobs are not timing out | 21:53 |
| guilhermesp | http://grafana.openstack.org/d/4-pSHcImz/openstack-ansible?orgId=1&from=now-6h&to=now&refresh=30s&panelId=10&fullscreen | 21:55 |
| jrosser | the most recent centos jobs in the integrated jobs have timed out in tempest rather than failed | 22:00 |
| jrosser | that seems to be a bit of progress | 22:00 |
| jamesdenton | are you seeing "waiting for privilege escalation prompt" or something else? | 22:01 |
| jrosser | the ones that timed out most recently didnt do that | 22:03 |
| jrosser | for the openstack-ansible repo anyway, i've just rechecked a bunch to get more examples | 22:04 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron stable/queens: Fix linuxbridge agent extensions https://review.openstack.org/620058 | 22:05 |
| guilhermesp | as Im watching zuul results for my, seems that only one centos job is going to timeout, but has finished the tempest execution | 22:06 |
| guilhermesp | seems really slow to run the log collection script | 22:07 |
| *** spatel has quit IRC | 22:08 | |
| jrosser | i wonder if it's doing a bunch of work on shared storage that should really be done in /dev/shm | 22:09 |
| *** aedc has joined #openstack-ansible | 22:29 | |
| *** rodolof has quit IRC | 22:37 | |
| *** vnogin has joined #openstack-ansible | 22:58 | |
| *** cshen has quit IRC | 22:59 | |
| *** vnogin has quit IRC | 23:02 | |
| *** spatel has joined #openstack-ansible | 23:04 | |
| *** hamzaachi__ has joined #openstack-ansible | 23:08 | |
| *** lbragstad has quit IRC | 23:08 | |
| *** hamzaachi_ has quit IRC | 23:09 | |
| *** spatel has quit IRC | 23:09 | |
| *** lbragstad has joined #openstack-ansible | 23:09 | |
| *** aedc has quit IRC | 23:12 | |
| *** DanyC has joined #openstack-ansible | 23:16 | |
| *** ThiagoCMC has quit IRC | 23:17 | |
| *** DanyC has quit IRC | 23:20 | |
| *** ThiagoCMC has joined #openstack-ansible | 23:27 | |
| redkrieg | hey jamesdenton thanks for the tip. I tried changing the openvswitch agent to use br-vlan instead but no luck there. the eth12 veth device that's tied in to br-provider is bridged to enp4s0f0 though, so I think that might be okay. I get normal connectivity, just not the dhcp stuff that should be coming in from neutron | 23:27 |
| *** hamzaachi__ has quit IRC | 23:32 | |
| ThiagoCMC | Hey guys, how to share one single Horizon with multiple OSA deployments (one complete OSA per Region), the "Federation" is basically a big LDAP that is already reachable across Regions, so, same login on each individual Horizon. | 23:41 |
| *** timburke has joined #openstack-ansible | 23:46 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!