| *** macza has quit IRC | 00:03 | |
| *** joshualyle has joined #openstack-ansible | 00:25 | |
| *** joshualyle has quit IRC | 00:30 | |
| *** hwoarang has quit IRC | 00:30 | |
| *** hwoarang has joined #openstack-ansible | 00:31 | |
| *** hwoarang has quit IRC | 00:41 | |
| *** hwoarang has joined #openstack-ansible | 00:42 | |
| *** joshualyle has joined #openstack-ansible | 00:58 | |
| *** kplant has quit IRC | 00:58 | |
| *** hwoarang has quit IRC | 01:00 | |
| *** hwoarang has joined #openstack-ansible | 01:00 | |
| *** markvoelker has quit IRC | 01:05 | |
| *** hwoarang has quit IRC | 01:13 | |
| *** hwoarang has joined #openstack-ansible | 01:14 | |
| *** BjoernT has quit IRC | 01:24 | |
| *** sreejithp has joined #openstack-ansible | 01:27 | |
| *** sreejithp has quit IRC | 01:31 | |
| *** hwoarang has quit IRC | 01:33 | |
| *** hwoarang has joined #openstack-ansible | 01:33 | |
| *** hwoarang has quit IRC | 01:47 | |
| *** hwoarang has joined #openstack-ansible | 01:47 | |
| *** BjoernT has joined #openstack-ansible | 01:51 | |
| *** hwoarang has quit IRC | 01:58 | |
| *** hwoarang has joined #openstack-ansible | 01:59 | |
| *** hwoarang has quit IRC | 02:04 | |
| *** hwoarang has joined #openstack-ansible | 02:04 | |
| *** udesale has quit IRC | 02:15 | |
| *** hwoarang has quit IRC | 02:32 | |
| *** hwoarang has joined #openstack-ansible | 02:32 | |
| *** joshualyle has quit IRC | 02:38 | |
| *** joshualyle has joined #openstack-ansible | 02:47 | |
| *** hwoarang has quit IRC | 02:48 | |
| *** hwoarang has joined #openstack-ansible | 02:49 | |
| *** joshualyle has quit IRC | 02:52 | |
| *** hwoarang has quit IRC | 03:03 | |
| *** hwoarang has joined #openstack-ansible | 03:03 | |
| *** markvoelker has joined #openstack-ansible | 03:06 | |
| *** jamesdenton has quit IRC | 03:20 | |
| *** hwoarang has quit IRC | 03:21 | |
| *** hwoarang has joined #openstack-ansible | 03:21 | |
| *** joshualyle has joined #openstack-ansible | 03:28 | |
| *** joshualyle has quit IRC | 03:30 | |
| *** djhankb has quit IRC | 03:33 | |
| *** shyamb has joined #openstack-ansible | 03:33 | |
| *** markvoelker has quit IRC | 03:40 | |
| *** pcaruana has joined #openstack-ansible | 04:22 | |
| *** joshualyle has joined #openstack-ansible | 04:29 | |
| *** ivve has quit IRC | 04:32 | |
| *** joshualyle has quit IRC | 04:34 | |
| *** markvoelker has joined #openstack-ansible | 04:37 | |
| *** joshualyle has joined #openstack-ansible | 04:38 | |
| *** hwoarang has quit IRC | 04:38 | |
| *** hwoarang has joined #openstack-ansible | 04:40 | |
| *** joshualyle has quit IRC | 04:42 | |
| *** shyamb has quit IRC | 04:48 | |
| *** raukadah is now known as chandankumar | 04:52 | |
| *** hwoarang has quit IRC | 05:05 | |
| *** hwoarang has joined #openstack-ansible | 05:06 | |
| *** shyamb has joined #openstack-ansible | 05:08 | |
| *** markvoelker has quit IRC | 05:10 | |
| *** hwoarang has quit IRC | 05:16 | |
| *** hwoarang has joined #openstack-ansible | 05:17 | |
| *** hwoarang has quit IRC | 05:22 | |
| *** hwoarang has joined #openstack-ansible | 05:22 | |
| *** hwoarang has quit IRC | 05:35 | |
| *** hwoarang has joined #openstack-ansible | 05:35 | |
| *** ivve has joined #openstack-ansible | 05:38 | |
| *** BjoernT has quit IRC | 05:40 | |
| *** shyamb has quit IRC | 06:02 | |
| *** kopecmartin|off is now known as kopecmartin | 06:04 | |
| *** shyamb has joined #openstack-ansible | 06:05 | |
| *** markvoelker has joined #openstack-ansible | 06:07 | |
| *** hwoarang has quit IRC | 06:11 | |
| *** hwoarang has joined #openstack-ansible | 06:12 | |
| *** shyamb has quit IRC | 06:29 | |
| *** cshen has joined #openstack-ansible | 06:31 | |
| *** shyamb has joined #openstack-ansible | 06:31 | |
| *** yolanda has quit IRC | 06:38 | |
| *** markvoelker has quit IRC | 06:40 | |
| openstackgerrit | Chandan Kumar (raukadah) proposed openstack/openstack-ansible-os_tempest master: Update openstack.org -> opendev.org https://review.opendev.org/654942 | 06:43 |
|---|---|---|
| *** macza has joined #openstack-ansible | 06:45 | |
| *** macza has quit IRC | 06:50 | |
| _janno | I'd like to split network_hosts into agents and servers. how can I achieve this? | 06:51 |
| *** shyam89 has joined #openstack-ansible | 07:05 | |
| *** shyamb has quit IRC | 07:07 | |
| *** starborn has joined #openstack-ansible | 07:08 | |
| fnpanic | morning | 07:09 |
| *** jhesketh has quit IRC | 07:11 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible master: Imported Translations from Zanata https://review.opendev.org/661551 | 07:17 |
| *** shyam89 has quit IRC | 07:24 | |
| *** shyamb has joined #openstack-ansible | 07:24 | |
| *** jhesketh has joined #openstack-ansible | 07:29 | |
| *** logan- has quit IRC | 07:33 | |
| *** aedc has joined #openstack-ansible | 07:33 | |
| *** tosky has joined #openstack-ansible | 07:35 | |
| *** logan- has joined #openstack-ansible | 07:36 | |
| *** markvoelker has joined #openstack-ansible | 07:37 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_neutron master: Remove name2int https://review.opendev.org/656761 | 07:37 |
| *** shyam89 has joined #openstack-ansible | 07:43 | |
| *** shyamb has quit IRC | 07:43 | |
| *** Emine has joined #openstack-ansible | 07:58 | |
| *** shyam89 has quit IRC | 08:04 | |
| *** markvoelker has quit IRC | 08:10 | |
| *** priteau has joined #openstack-ansible | 08:57 | |
| *** shyamb has joined #openstack-ansible | 09:05 | |
| *** markvoelker has joined #openstack-ansible | 09:07 | |
| *** taseer has joined #openstack-ansible | 09:08 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-repo_server master: Allow remote detection of repo sync status https://review.opendev.org/660788 | 09:11 |
| CeeMac | morning | 09:17 |
| _janno | In other words: How can I run the neutron_server_containers on other hosts than the network_hosts? | 09:18 |
| CeeMac | _janno, somone was asking about this in the channel a short while ago. I'll see if I can find the answer | 09:20 |
| _janno | CeeMac: Thank you, I'll have a look, too. | 09:21 |
| CeeMac | i remember you needed to create a new skeleton | 09:21 |
| jrosser | yes this was asked within the last week i think and cloudnull posted a snippet that would do it | 09:22 |
| _janno | ok, that's what I am already trying. | 09:22 |
| _janno | but so far I've failed | 09:22 |
| jrosser | _janno: here you go http://eavesdrop.openstack.org/irclogs/%23openstack-ansible/%23openstack-ansible.2019-05-21.log.html#t2019-05-21T21:10:27 | 09:24 |
| CeeMac | jrosser: is there a search facility on the eavsdrop? | 09:25 |
| jrosser | not that i know of - i brute forced that :/ | 09:26 |
| _janno | jrosser: thanks a lot | 09:26 |
| *** shyamb has quit IRC | 09:26 | |
| CeeMac | good work | 09:26 |
| *** sreejithp has joined #openstack-ansible | 09:28 | |
| _janno | I am wondering if and then why it is unusual to run neutron agent and server on different hosts. | 09:28 |
| CeeMac | _janno: I was actually wondering this myself earlier today | 09:30 |
| *** taseer has quit IRC | 09:31 | |
| *** sreejithp has quit IRC | 09:33 | |
| CeeMac | i guess it all comes down to the individual environment | 09:34 |
| *** shyamb has joined #openstack-ansible | 09:34 | |
| CeeMac | was pondering whether moving the neutron-server container to the control nodes would be more efficient, or running neutron-server as metal on the network nodes. Or just leave it, presumably theres a good reason why it is that way :) | 09:35 |
| jrosser | i don't really expect there is a good reason that the neutron server 'sticks' to the network nodes | 09:39 |
| *** markvoelker has quit IRC | 09:40 | |
| jrosser | i suspect the most consistent pattern would be for those containers to stay on the infra nodes when network nodes are separated out | 09:40 |
| *** hamzaachi has joined #openstack-ansible | 09:44 | |
| jrosser | CeeMac: are you considering moving the neutron servers back to your infra nodes? | 09:45 |
| *** jcosmao has joined #openstack-ansible | 09:48 | |
| *** shyamb has quit IRC | 09:52 | |
| *** shyam89 has joined #openstack-ansible | 09:52 | |
| CeeMac | jrosser: I was pondering it yeah. | 10:14 |
| CeeMac | jrosser: probably the controllers with the other api containers | 10:15 |
| CeeMac | just trying to work out how much effort that would be and if its worth it | 10:16 |
| *** emine__ has joined #openstack-ansible | 10:20 | |
| *** Emine has quit IRC | 10:22 | |
| jrosser | CeeMac: the effort is probably relatively small if the paste from last week is good | 10:30 |
| *** markvoelker has joined #openstack-ansible | 10:37 | |
| CeeMac | jrosser: yeah, just need to clean the inventory and redeploy the containers on top of that i guess | 10:40 |
| jrosser | yes i think that should do it | 10:40 |
| CeeMac | i guess most of the work is done by the agents, so the additional overhead on the controllers shouldn't be much | 10:41 |
| jrosser | right, and it might be a good thing to remove the servers from any potential load/self inflicted DOS on the network nodes | 10:41 |
| CeeMac | yeah, good point | 10:41 |
| *** aedc has quit IRC | 10:42 | |
| CeeMac | would probably be good for people who want to run collapsed compute/network agent nodes as well | 10:42 |
| CeeMac | which I might need to do in one of the other DCs | 10:42 |
| CeeMac | running a bit tight for time. lost most of yesterday troubleshooting one of the nodes not coming back online after a reboot | 10:43 |
| CeeMac | turns out that usb-storage is disabled by the security-hardening playbook. | 10:44 |
| CeeMac | oddly enough, embedded SD modules require that to load, and if your /boot partition goes missing the server doesn't like it | 10:44 |
| *** k4bi has quit IRC | 10:44 | |
| CeeMac | that was fun to troubleshoot | 10:44 |
| *** shyam89 has quit IRC | 10:47 | |
| _janno | well, another related question: So far we had the network stuff running on the control nodes. Now I would like to move it to dedicated network nodes. I've added the neutron_server.yml from above. What do I have to remove from the inventory now? What's about the running neutron processes on the control nodes? | 10:48 |
| CeeMac | _janno, you want to move just the agents off the control nodes? | 10:53 |
| _janno | CeeMac: yes, right | 10:54 |
| CeeMac | _janno: from that skel, I think you'd need to specify your network node IPs under the 'network_containers' host group, and the control nodes under the 'network_hosts' host group in your openstack_user_config.yml | 10:57 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_tempest master: Default network segmentation ID correctly https://review.opendev.org/659877 | 10:59 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_tempest master: [redhat] Update variables to be version agnostic https://review.opendev.org/658262 | 11:02 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_tempest master: Add support for aarch64 images https://review.opendev.org/620032 | 11:02 |
| *** markvoelker has quit IRC | 11:11 | |
| *** dave-mccowan has joined #openstack-ansible | 11:14 | |
| *** shyamb has joined #openstack-ansible | 11:19 | |
| _janno | CeeMac: okay, but I am wondering if osa takes care of the processes which are currently running on the control nodes | 11:19 |
| *** kplant has joined #openstack-ansible | 11:24 | |
| CeeMac | ah | 11:25 |
| CeeMac | _janno: right, yeah, good point | 11:26 |
| CeeMac | how does one remove the metal config. for containers you can destroy containers | 11:26 |
| CeeMac | not sure if that covers metal agents | 11:28 |
| CeeMac | let me poke about a bit | 11:28 |
| jrosser | you can probably migrate the routers to the new nodes, and disable the service | 11:29 |
| CeeMac | jrosser: I guess you'd need to manually uninstall the agents from the controllers? | 11:30 |
| *** macza has joined #openstack-ansible | 11:39 | |
| *** macza has quit IRC | 11:43 | |
| *** ansmith has quit IRC | 11:47 | |
| *** mkuf_ has joined #openstack-ansible | 11:56 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Update to ceph-ansible 4.0.0 https://review.opendev.org/656503 | 11:56 |
| kplant | has anyone thought of a clever way to not waste a floating IP on each compute node when using DVR? | 11:59 |
| *** mkuf has quit IRC | 12:00 | |
| *** electrofelix has joined #openstack-ansible | 12:02 | |
| CeeMac | kplant: not entirely sure thats a thing | 12:02 |
| CeeMac | its listed as one of the 'cons' of using DVR i believe | 12:02 |
| kplant | yeah it is, it's just so wasteful that i'm trying to spend cycles on out-smarting or out-dumbing it | 12:04 |
| kplant | if i have 1k compute nodes, i don't think burning 1k public IPs is okay | 12:04 |
| CeeMac | haha, sound fun! Let me know if you come up with anything | 12:04 |
| CeeMac | ouch | 12:04 |
| CeeMac | yeah, thats going to cost some | 12:04 |
| CeeMac | you could look at ovn as a alternative maybe? I think it handles the distributed routing differently | 12:05 |
| *** markvoelker has joined #openstack-ansible | 12:06 | |
| CeeMac | you couldn't even get around it using NAT as its the FIP namespace running on the computer nodes so it'd need 1-1 mapping anyhow | 12:06 |
| kplant | i think static nats could work, but it would be a nightmare to manage | 12:07 |
| kplant | real external ip <-> floating ip <-> private vxlan ip | 12:08 |
| CeeMac | you'd still need 1 per compute node though? | 12:08 |
| CeeMac | for the router instance | 12:08 |
| kplant | one fip yes, but if you make those rfc1918 or squat then who cares? :D | 12:08 |
| CeeMac | hmm, then you need to manage routing and nat in front, more hastle :) | 12:09 |
| kplant | exactly | 12:09 |
| CeeMac | you could not DVR? lol | 12:09 |
| kplant | yep :| | 12:09 |
| nowster | IPv6 ? ;-) | 12:09 |
| CeeMac | pfft, noone uses that :p | 12:10 |
| kplant | i work for an isp so... we use ipv6 the least out of everybody | 12:10 |
| kplant | makes sense! | 12:10 |
| CeeMac | or you'd need a 4-6 tunnel in front for non ipv6 endpoints to connect in to VMs | 12:10 |
| CeeMac | yeah, IPv6 onthe edge can't be a thing until the telcos start rolling out IPv6 in homes/businesses | 12:11 |
| kplant | i think all of our resi supports ipv6, with dhcpv6 (why!?!?!) and our fia has been silently including /48s for a while | 12:11 |
| kplant | but we don't use it internally (at all) | 12:12 |
| CeeMac | ah, you must be outside uk then? | 12:12 |
| kplant | yeah, us | 12:12 |
| CeeMac | figures | 12:12 |
| nowster | In the UK, all the large ISPs support v6 with the exception of the cable ISP (Virgin). | 12:12 |
| CeeMac | nowster, support or use? | 12:12 |
| nowster | support | 12:13 |
| CeeMac | thats the catch | 12:13 |
| nowster | Rumour has it that Virgin Media will provide native IPv6 with CGNAT for IPv4. The gamers aren't pleased. | 12:13 |
| *** priteau has quit IRC | 12:13 | |
| *** priteau has joined #openstack-ansible | 12:14 | |
| CeeMac | nothing pleases gamers :D | 12:14 |
| nowster | Indeed. They want latencies less than what the speed of light allows. | 12:15 |
| kplant | then you get the ones who are "smart" and calculate the disatnce from them to the server and divide it by c | 12:15 |
| CeeMac | but why can't you make it faster than light? etc | 12:15 |
| kplant | index of refraction is not a thing, round trip time is not a thing | 12:15 |
| CeeMac | science has no place in logic dag namit! | 12:16 |
| CeeMac | i'm closer so it should be faster. make it so! | 12:16 |
| kplant | why are there hops on the traceroute!? i should have a direct connection | 12:17 |
| kplant | your network sucks | 12:17 |
| jrosser | kplant: i think you can put an rfc1918 address on the computes | 12:17 |
| jrosser | and have that as a hop between your computes and the next router with a real IP | 12:17 |
| kplant | oooooooooo | 12:18 |
| kplant | and route through | 12:18 |
| kplant | that's an interesting idea | 12:18 |
| jrosser | so as an example, at home on my DSL my next hop is a 172.29.blah even though i have a real IP on my router | 12:18 |
| kplant | at that point it's not pat/nat, it's just routing | 12:18 |
| jrosser | exactly | 12:19 |
| *** shyamb has quit IRC | 12:19 | |
| CeeMac | oh, they do that for MPLS and IPVPN as well for the internal hops between routers | 12:19 |
| jrosser | i was asking the neutron folks about exactly this at the summit and it was their answer | 12:19 |
| kplant | i'm hesitant to call that a great idea because it's so simple | 12:20 |
| kplant | and i'm afraid to get excited | 12:20 |
| *** mkuf_ has quit IRC | 12:21 | |
| kplant | how do you handle ingress? | 12:21 |
| kplant | sounds like you'd need a routing protocol | 12:21 |
| CeeMac | would you need to pair this with the BGPaaS plugin? | 12:22 |
| kplant | i'm thinking the real router and virtual routers need to exchange routes somehow | 12:22 |
| *** shyamb has joined #openstack-ansible | 12:23 | |
| kplant | the virtual routers can all statically sent to the real gateway | 12:23 |
| kplant | but the real router is going to have to know which virtual gateway to route through dynamically | 12:23 |
| kplant | or am i overcomplicating this | 12:24 |
| CeeMac | i think you're probably right | 12:24 |
| *** mkuf_ has joined #openstack-ansible | 12:24 | |
| CeeMac | its probably worth reviewing the bgpaas docs to see if it meets the remit | 12:25 |
| jrosser | sounds right | 12:26 |
| CeeMac | as its ingress, there would need to be a way for the edge router to know which VR IP to send the packet to for the particular FIP in question | 12:27 |
| CeeMac | and i don't think proxy arp would be an option so it would need to be a routing protocol | 12:27 |
| CeeMac | not sure how DVR sits with the various *aaS plugins though. | 12:28 |
| kplant | yeah i believe the neutron-bgp service does self service network routing? | 12:28 |
| CeeMac | i believe that is its principal focus, not sure if it can be adapted to provider routing though | 12:29 |
| jrosser | there is some stuff here anout BGP and DVR https://docs.openstack.org/neutron/rocky/admin/config-bgp-dynamic-routing.html | 12:30 |
| jrosser | *about | 12:30 |
| *** aedc has joined #openstack-ansible | 12:30 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/stein: Use stable ARA setup interface for defining the callback location https://review.opendev.org/661313 | 12:30 |
| CeeMac | hmm, it does seem focused on self-service networks | 12:31 |
| kplant | i'm reading the dvr section of that | 12:31 |
| *** chason has joined #openstack-ansible | 12:31 | |
| jrosser | "For floating IP addresses, the BGP speaker advertises the floating IP agent gateway on the corresponding compute node as the next-hop IP address" <- sounds like what you need | 12:31 |
| CeeMac | i just got that far too | 12:32 |
| CeeMac | looks interesting actually | 12:33 |
| CeeMac | I'm going to bookmark this for later thought | 12:33 |
| *** shyamb has quit IRC | 12:34 | |
| jrosser | kplant: i hope that gives you some ideas - i'm not sure there is much alternative if you want to do DVR and conserve address space with a large number of computes | 12:35 |
| kplant | yeah for sure and ebgp is about as simple as it gets | 12:35 |
| kplant | thank you | 12:35 |
| CeeMac | looks like you've got some fun ahead kplant :) | 12:37 |
| jrosser | it has the nice side effect that your compute nodes and intermediate routers aren't actually routable from outside your AS | 12:38 |
| kplant | yeah i can't wait to run into timing issues because palo alto _sucks_ at bgp | 12:38 |
| CeeMac | jrosser: i like that side effect | 12:38 |
| *** ansmith has joined #openstack-ansible | 12:41 | |
| mgariepy | can i have one last push on this one please: https://review.opendev.org/#/c/659277/ | 12:43 |
| kplant | might be able to use the entire cidr for fip too, doesn't sound like you need to burn a gw ip | 12:44 |
| CeeMac | I'd be interested to hear how you get on kplant | 12:47 |
| kplant | i have quite the dilemma, i have to also think about making this supportable for the rest of my team | 12:48 |
| kplant | absolutely none of them have a network engineering background | 12:48 |
| kplant | this might be a 'cool poc' in the lab that gets pushed back forever | 12:48 |
| jrosser | it's difficult isnt it, becasue a lot of the coolest networking stuffs for datacenter seems to be in some way involving L3 and BGP | 12:53 |
| kplant | right?? | 12:54 |
| _janno | sure https://imgflip.com/i/32279u | 12:56 |
| CeeMac | _janno lol | 12:57 |
| kplant | everything should be a /32 or /128 | 12:57 |
| kplant | routes for all | 12:57 |
| CeeMac | jrosser: the modern utopian dream for the DC seems to be pure L3 and dispatch with L2 | 12:58 |
| CeeMac | solves so many issues, but introduces other challenges | 12:58 |
| jrosser | kplant: also look at this where the p2p are made with ipv6 for a v4 network https://docs.cumulusnetworks.com/display/ROH/Routing+on+the+Host | 12:58 |
| CeeMac | hmm, could you do the edge/fip routing with calico? | 12:59 |
| CeeMac | or can that not run side by side | 12:59 |
| *** dave-mccowan has quit IRC | 13:00 | |
| kplant | sdn and policy based routing can probably solve it | 13:00 |
| CeeMac | yeah, thats why i mentioned ovn as a possible option. that brings its own complexities though | 13:00 |
| CeeMac | unless you have an sdn fabric already? | 13:01 |
| kplant | nothing fancy, just ovs | 13:01 |
| kplant | not using it for anything other than vxlan | 13:01 |
| CeeMac | thats what I'm in the process of putting in | 13:02 |
| kplant | i could make the juniper overlords happy and try tungsten | 13:02 |
| _janno | what's the advantage of ovs over linux-bridge in case of vxlan? | 13:02 |
| CeeMac | actually, it isnt. I'm using it for provider networks | 13:02 |
| kplant | _janno: i don't think there is one, anymore | 13:02 |
| CeeMac | juniper guys at the summit were excited about tungsten | 13:03 |
| _janno | great, I'd like to stick with linux-bridge :) | 13:03 |
| CeeMac | I'm tempted, but seems like a lot of effort | 13:03 |
| CeeMac | _janno: unless you want DVR or other advances features, no need to leave lxb | 13:04 |
| kplant | linux bridges probably perform better, less overhead | 13:06 |
| kplant | and it's right in the kernel | 13:06 |
| *** zul has joined #openstack-ansible | 13:07 | |
| nowster | LXB is for when you can assign a block of VLANs yourself; VxLAN is for when you can't. | 13:12 |
| kplant | lxb can do vxlan as well | 13:13 |
| *** BjoernT has joined #openstack-ansible | 13:19 | |
| *** priteau has quit IRC | 13:20 | |
| *** tosky__ has joined #openstack-ansible | 13:21 | |
| *** tosky is now known as Guest72073 | 13:21 | |
| *** tosky__ is now known as tosky | 13:21 | |
| openstackgerrit | Andreas Jaeger proposed x/monitorstack master: Fix docs building https://review.opendev.org/661633 | 13:22 |
| *** sreejithp has joined #openstack-ansible | 13:26 | |
| *** BjoernT has quit IRC | 13:32 | |
| *** sreejithp has quit IRC | 13:32 | |
| *** BjoernT has joined #openstack-ansible | 13:41 | |
| *** starborn has quit IRC | 14:08 | |
| *** hamzaachi_ has joined #openstack-ansible | 14:10 | |
| *** hamzaachi has quit IRC | 14:14 | |
| openstackgerrit | kourosh vivan proposed openstack/openstack-ansible master: rgw keystone setup fix python interpreter https://review.opendev.org/661006 | 14:19 |
| *** hamzaachi__ has joined #openstack-ansible | 14:20 | |
| *** shyamb has joined #openstack-ansible | 14:22 | |
| *** hamzaachi_ has quit IRC | 14:22 | |
| *** hamzaachi__ has quit IRC | 14:24 | |
| *** hamzaachi has joined #openstack-ansible | 14:24 | |
| *** chandankumar is now known as raukadah | 14:25 | |
| *** hamzaachi has quit IRC | 14:26 | |
| *** hamzaachi has joined #openstack-ansible | 14:26 | |
| *** hamzaachi has quit IRC | 14:28 | |
| *** hamzaachi has joined #openstack-ansible | 14:30 | |
| *** shyamb has quit IRC | 14:33 | |
| *** shyamb has joined #openstack-ansible | 14:34 | |
| *** hamzaachi_ has joined #openstack-ansible | 14:35 | |
| *** dasp has quit IRC | 14:36 | |
| *** hamzaachi has quit IRC | 14:37 | |
| *** hamzaachi_ has quit IRC | 14:42 | |
| *** hamzaachi_ has joined #openstack-ansible | 14:42 | |
| *** hamzaachi_ has quit IRC | 14:46 | |
| *** hamzaachi_ has joined #openstack-ansible | 14:46 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/stein: Correct os-cinder-install task serialisation https://review.opendev.org/661649 | 14:53 |
| openstackgerrit | Merged openstack/openstack-ansible stable/rocky: Correct os-cinder-install task serialisation https://review.opendev.org/661650 | 14:53 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Correct os-cinder-install task serialisation https://review.opendev.org/661651 | 14:53 |
| *** hamzaachi_ has quit IRC | 14:53 | |
| *** shyam89 has joined #openstack-ansible | 14:57 | |
| *** devx has quit IRC | 14:57 | |
| *** devx has joined #openstack-ansible | 14:58 | |
| *** udesale has joined #openstack-ansible | 15:00 | |
| *** shyamb has quit IRC | 15:00 | |
| *** ivve has quit IRC | 15:03 | |
| *** djhankb has joined #openstack-ansible | 15:08 | |
| *** hamzaachi has joined #openstack-ansible | 15:10 | |
| *** shyam89 has quit IRC | 15:14 | |
| *** tosky has quit IRC | 15:16 | |
| *** shyamb has joined #openstack-ansible | 15:17 | |
| *** hamzy has quit IRC | 15:18 | |
| *** cshen has quit IRC | 15:23 | |
| openstackgerrit | Andreas Jaeger proposed x/monitorstack master: Update zuul jobs https://review.opendev.org/661614 | 15:38 |
| *** macza has joined #openstack-ansible | 15:41 | |
| *** shyam89 has joined #openstack-ansible | 15:43 | |
| *** shyamb has quit IRC | 15:44 | |
| *** macza has quit IRC | 15:44 | |
| *** macza has joined #openstack-ansible | 15:45 | |
| *** hamzy has joined #openstack-ansible | 15:49 | |
| *** gshippey has joined #openstack-ansible | 15:51 | |
| *** hamzaachi_ has joined #openstack-ansible | 15:53 | |
| gshippey | Anyone free to review a oneliner patch https://review.opendev.org/#/c/660779/ ? | 15:53 |
| *** hamzaachi has quit IRC | 15:55 | |
| evrardjp | gshippey: aren't those in a venv instead ? | 15:56 |
| *** kopecmartin is now known as kopecmartin|off | 15:58 | |
| gshippey | @evrardjp the variable we are adding to gets used to install the pip packages to the venv, see https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/utility-install.yml#L124 | 15:58 |
| *** shyam89 has quit IRC | 15:58 | |
| gshippey | Maybe i need to be clearer in the commit message | 15:58 |
| jrosser | logan-: should this be conditional on the unbound group being length>0 ? https://github.com/openstack/openstack-ansible/blob/master/playbooks/setup-infrastructure.yml#L16 | 16:01 |
| jrosser | with the group empty it currently targets no hosts for the server, but still everything except 'unbound' for the clients | 16:01 |
| evrardjp | gshippey: I would remove this var from group vars I think. | 16:01 |
| logan- | hmm jrosser.. https://github.com/openstack/openstack-ansible/blob/c012df25aba9c50f03f81285063e1a78e8d6d8f6/playbooks/unbound-install.yml#L42-L43 should prevent that | 16:01 |
| logan- | well, the playbook would still scope the clients, except they should all skip | 16:02 |
| jrosser | logan-: ahha i see --list-hosts isnt smart enough to understand that | 16:03 |
| *** shyamb has joined #openstack-ansible | 16:03 | |
| logan- | oh gotcha | 16:03 |
| jrosser | was just double checking "what to run for adding a compute node" | 16:03 |
| evrardjp | jrosser: don't we have up to date docs for that ? ;) | 16:04 |
| jrosser | evrardjp: well this https://docs.openstack.org/openstack-ansible/latest/admin/scale-environment.html#add-a-compute-host | 16:05 |
| jrosser | but there seem to be residual parts of setup-infrastructure which touch compute nodes | 16:05 |
| evrardjp | which ones? Just curious there. | 16:06 |
| jrosser | or perhaps that should be "as reported by openstack-ansible setup-infrastructure.yml --list-hosts" | 16:06 |
| evrardjp | aaaaah ;) | 16:07 |
| evrardjp | static analysis of ansible runs is something that severely lacks | 16:07 |
| jrosser | so this | 16:08 |
| evrardjp | I think Adam wrote something for this | 16:08 |
| jrosser | play #8 (hosts:all_containers): Configure all nodes to use the repo container for python/apt packagesTAGS: [repo-config,always] | 16:08 |
| jrosser | pattern: [u'hosts:all_containers'] | 16:08 |
| jrosser | that appears to be part of setup-infrastructure that targets * | 16:08 |
| evrardjp | we can probably make that smarter then | 16:09 |
| * jrosser raised an eyebrow at that | 16:09 | |
| jrosser | yeah, it doesnt feel right | 16:09 |
| evrardjp | I am on my phone. If you have a link to the playbook it would make me earn precious minutes | 16:09 |
| *** gyee has joined #openstack-ansible | 16:20 | |
| jrosser | evrardjp: this is repo-use, on rocky | 16:22 |
| *** ivve has joined #openstack-ansible | 16:22 | |
| jrosser | https://github.com/openstack/openstack-ansible/blob/stable/rocky/playbooks/repo-use.yml#L17 | 16:22 |
| *** admin0 has quit IRC | 16:22 | |
| evrardjp | I see where this comes from | 16:22 |
| evrardjp | I doubt RAX still cares. | 16:23 |
| evrardjp | in other words, it's probably possible to refactor | 16:23 |
| evrardjp | oh I remember now | 16:24 |
| evrardjp | chicken and egg thing :) | 16:24 |
| evrardjp | you needed the repo build to finish to then point to it. | 16:25 |
| mnaser | Yes. You can't repo use first or else the world will break | 16:25 |
| evrardjp | repo build is now away, we can move that in a different role | 16:25 |
| mnaser | It's no longer a thing with venv build | 16:26 |
| evrardjp | yup | 16:26 |
| noonedeadpunk | wow, isn't it meeting time? | 16:26 |
| *** shyamb has quit IRC | 16:26 | |
| *** emine__ has quit IRC | 16:27 | |
| *** shyamb has joined #openstack-ansible | 16:27 | |
| *** prometheanfire has quit IRC | 16:28 | |
| *** hamzaachi_ has quit IRC | 16:29 | |
| *** prometheanfire has joined #openstack-ansible | 16:29 | |
| evrardjp | it is | 16:29 |
| evrardjp | for still 30 minutes | 16:29 |
| evrardjp | :) | 16:29 |
| noonedeadpunk | ok, I see:) | 16:30 |
| evrardjp | I don't have much to report on suse support nor releases | 16:31 |
| noonedeadpunk | I probably have nothing to share, just realized what time is now | 16:31 |
| evrardjp | haha | 16:31 |
| noonedeadpunk | do we have some agenda (read eatherpad) of things that should be done before release? | 16:32 |
| jrosser | if its meeting time i would ask for a bit of help with reviews | 16:32 |
| evrardjp | noonedeadpunk: there is docs that I need to update, per branch. Want to help? | 16:33 |
| evrardjp | it's now super easy | 16:33 |
| noonedeadpunk | evrardjp: If there's smth I can help with:) | 16:34 |
| evrardjp | there sure is :) | 16:34 |
| *** shyam89 has joined #openstack-ansible | 16:36 | |
| guilhermesp | hum there are two things actually I'd like to discuss | 16:37 |
| *** shyamb has quit IRC | 16:37 | |
| guilhermesp | we kinda of have already a refactor pattern https://review.opendev.org/#/q/topic:osa/db-refactor . We just had a discussion here on sunday about run_once in keystone | 16:38 |
| guilhermesp | I changed it to use conditionals, but maybe there is something else to do to make the functional jobs to work https://review.opendev.org/#/c/657036/ | 16:38 |
| openstackgerrit | Georgina Shippey proposed openstack/openstack-ansible stable/rocky: Utility container is missing the shade library https://review.opendev.org/661812 | 16:40 |
| *** shyam89 has quit IRC | 16:41 | |
| logan- | these are looking good guilhermesp. any idea why those 2 jobs are failing on keystone? seems like its failing to db sync? | 16:41 |
| logan- | i'm not seeing any logging from keystone-manage in the gate logs so it is hard to tell what's going on :/ | 16:42 |
| nowster | jrosser: thanks for rebasing it again. | 16:44 |
| jrosser | no worries - it passed i think? | 16:44 |
| nowster | yep! | 16:44 |
| jrosser | \o/ | 16:45 |
| guilhermesp | logan-: I think I have local env that I was testing thing out | 16:45 |
| guilhermesp | let me see | 16:45 |
| jrosser | nowster: is there anything else outstanding for aarch64? | 16:46 |
| jrosser | well, apart from there still being no practical way to run a CI job..... | 16:47 |
| nowster | jrosser: and wheels? | 16:49 |
| jrosser | yes indeed - thats the major hurdle now i think | 16:49 |
| nowster | jrosser: sorry about the lack of servers... the rug was pulled out from under us on that | 16:49 |
| evrardjp | is there a problem for wheel build ? | 16:50 |
| jrosser | yes i gathered as much from the summit | 16:50 |
| jrosser | evrardjp: there are none for aarch64 | 16:50 |
| guilhermesp | logan-: yum seems that services is getting 503. I will dig into it eventually | 16:50 |
| evrardjp | maybe experience can be shared with loci there | 16:50 |
| evrardjp | there are some afaik | 16:50 |
| jrosser | evrardjp: if you know where that would be very useful | 16:51 |
| evrardjp | let's sync tomorrow. hrw in loci channel builds the requirements image for loci , IIRC the IRC nick | 16:53 |
| evrardjp | for arm | 16:53 |
| evrardjp | which means if it's published you can get that as a tar which contains the wheels. | 16:53 |
| evrardjp | all of those are the ones in u-c, but might differ from our requirements sha | 16:54 |
| evrardjp | just giving opportunities here | 16:55 |
| evrardjp | :) | 16:55 |
| *** aedc has quit IRC | 17:02 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible stable/rocky: Utility container venv is missing python libs for openstack ansible modules https://review.opendev.org/661812 | 17:05 |
| *** Weifan has joined #openstack-ansible | 17:12 | |
| *** irclogbot_1 has quit IRC | 17:17 | |
| *** irclogbot_0 has joined #openstack-ansible | 17:18 | |
| *** sreejithp has joined #openstack-ansible | 17:25 | |
| *** admin0 has joined #openstack-ansible | 17:44 | |
| *** electrofelix has quit IRC | 17:49 | |
| *** spatel has joined #openstack-ansible | 17:53 | |
| spatel | cloudnull: hey!! | 17:53 |
| spatel | how you doing ? | 17:54 |
| spatel | i came to find james d. but look like he is not online.. | 17:54 |
| *** Weifan has quit IRC | 17:57 | |
| *** Weifan has joined #openstack-ansible | 17:57 | |
| *** Weifan has quit IRC | 18:01 | |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-os_glance master: db_setup: refactor database setup to a common file https://review.opendev.org/657029 | 18:30 |
| BjoernT | Watch-out for https://bugs.launchpad.net/openstack-ansible/+bug/1830780 , Stein broke neutron multiple external networks | 18:34 |
| openstack | Launchpad bug 1830780 in openstack-ansible "Multiple external network support is broken in stein" [Undecided,New] | 18:34 |
| *** hwoarang has quit IRC | 18:41 | |
| *** hwoarang has joined #openstack-ansible | 18:41 | |
| *** Weifan has joined #openstack-ansible | 18:42 | |
| *** Weifan has quit IRC | 18:47 | |
| *** tstrul has joined #openstack-ansible | 19:11 | |
| *** mcarden has quit IRC | 19:16 | |
| tstrul | Hi all, I just finished upgrading my environment to queens after we were using ocata for a long time (ocata --> pike --> queens). after upgrading I can see heartbeat miss in rabbitmq logs in addition to messages similar to: "Timed out waiting for a reply to message ID" on some services. I have seen this bug: https://bugs.launchpad.net/oslo.messaging/+bug/1800957 and I understand problem is amqp<2.4.1 on all services. I am | 19:19 |
| openstack | Launchpad bug 1800957 in oslo.messaging (Ubuntu) "Upgrading to pike version causes rabbit timeouts with ssl" [Undecided,New] | 19:19 |
| tstrul | currently on 17.1.9 but I didnt see a fix on any version. how can i resolve this issue? is there an update on this issue or should i just disable TLS for now? | 19:19 |
| *** gshippey has quit IRC | 19:24 | |
| *** pcaruana has quit IRC | 19:30 | |
| *** Weifan has joined #openstack-ansible | 19:34 | |
| *** Weifan has quit IRC | 19:39 | |
| *** Weifan has joined #openstack-ansible | 19:41 | |
| *** BjoernT has quit IRC | 19:41 | |
| *** aedc has joined #openstack-ansible | 19:42 | |
| *** kplant has quit IRC | 19:44 | |
| *** tstrul has quit IRC | 19:48 | |
| jrosser | well how should we deal with that bug ^ - how are we supposed to discover / handle this https://review.opendev.org/#/c/638461 | 19:59 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: db_setup: refactor database setup to a common file https://review.opendev.org/657038 | 20:00 |
| *** aedc has quit IRC | 20:06 | |
| *** aedc has joined #openstack-ansible | 20:06 | |
| *** aedc has quit IRC | 20:10 | |
| *** aedc has joined #openstack-ansible | 20:10 | |
| *** aedc has quit IRC | 20:16 | |
| *** dasp has joined #openstack-ansible | 20:20 | |
| *** hamzy has quit IRC | 20:21 | |
| *** sreejithp_ has joined #openstack-ansible | 20:30 | |
| *** sreejithp has quit IRC | 20:31 | |
| _janno | Is anyone already running osa with kernel 5.0? | 20:31 |
| _janno | modprobe: FATAL: Module nf_conntrack_ipv6 not found | 20:32 |
| *** hwoarang has quit IRC | 20:37 | |
| *** hwoarang has joined #openstack-ansible | 20:38 | |
| *** dasp has quit IRC | 20:38 | |
| *** dasp has joined #openstack-ansible | 20:38 | |
| *** ansmith has quit IRC | 20:48 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest master: [redhat] Update variables to be version agnostic https://review.opendev.org/658262 | 20:49 |
| *** dave-mccowan has joined #openstack-ansible | 21:08 | |
| *** dave-mccowan has quit IRC | 21:23 | |
| *** udesale has quit IRC | 21:31 | |
| *** tosky has joined #openstack-ansible | 21:36 | |
| *** hwoarang has quit IRC | 21:58 | |
| *** hwoarang has joined #openstack-ansible | 21:59 | |
| *** spatel has quit IRC | 22:03 | |
| *** dave-mccowan has joined #openstack-ansible | 22:12 | |
| *** ansmith has joined #openstack-ansible | 22:16 | |
| *** Emine has joined #openstack-ansible | 22:21 | |
| *** Talion has joined #openstack-ansible | 22:24 | |
| *** emine__ has joined #openstack-ansible | 22:27 | |
| *** Emine has quit IRC | 22:29 | |
| *** dave-mccowan has quit IRC | 22:31 | |
| *** hamzaachi has joined #openstack-ansible | 22:42 | |
| *** hamzaachi has quit IRC | 22:46 | |
| *** hamzaachi has joined #openstack-ansible | 22:46 | |
| *** hamzaachi has quit IRC | 22:47 | |
| *** hamzaachi has joined #openstack-ansible | 22:48 | |
| *** emine__ has quit IRC | 22:51 | |
| *** hamzaachi has quit IRC | 22:59 | |
| *** hamzaachi has joined #openstack-ansible | 22:59 | |
| *** tosky has quit IRC | 23:04 | |
| *** spatel has joined #openstack-ansible | 23:10 | |
| *** dave-mccowan has joined #openstack-ansible | 23:13 | |
| *** spatel has quit IRC | 23:14 | |
| *** hamzaachi has quit IRC | 23:24 | |
| *** dave-mccowan has quit IRC | 23:33 | |
| *** macza has quit IRC | 23:36 | |
| *** jbadiapa has quit IRC | 23:37 | |
| *** jbadiapa has joined #openstack-ansible | 23:38 | |
| *** gyee has quit IRC | 23:40 | |
| *** dave-mccowan has joined #openstack-ansible | 23:46 | |
| *** dave-mccowan has quit IRC | 23:53 | |
| *** Weifan has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!