| *** salv-orlando has quit IRC | 00:03 | |
| *** chandrav has quit IRC | 00:27 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 00:49 | |
| *** markmcclain has quit IRC | 01:02 | |
| *** markmcclain has joined #openstack-neutron-ovn | 01:05 | |
| *** HenryG_ is now known as HenryG | 01:08 | |
| *** salv-orlando has quit IRC | 01:23 | |
| *** markmcclain has quit IRC | 01:43 | |
| *** markmcclain has joined #openstack-neutron-ovn | 01:45 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 01:45 | |
| *** yamamoto has joined #openstack-neutron-ovn | 01:54 | |
| *** yamamoto has quit IRC | 02:00 | |
| *** salv-orlando has quit IRC | 02:17 | |
| *** yamamoto has joined #openstack-neutron-ovn | 02:23 | |
| *** yamamoto has quit IRC | 02:28 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 02:42 | |
| *** salv-orlando has quit IRC | 03:26 | |
| *** yamamoto_ has joined #openstack-neutron-ovn | 03:33 | |
| *** gongysh_ has joined #openstack-neutron-ovn | 03:38 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 04:24 | |
| *** allan_h has quit IRC | 04:39 | |
| *** allan_h has joined #openstack-neutron-ovn | 04:39 | |
| *** allan_h has quit IRC | 04:56 | |
| *** salv-orlando has quit IRC | 05:06 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 05:17 | |
| *** numans has joined #openstack-neutron-ovn | 05:20 | |
| *** salv-orlando has quit IRC | 05:35 | |
| *** gongysh_ has quit IRC | 05:36 | |
| *** markmcclain has quit IRC | 06:13 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 06:13 | |
| *** markmcclain has joined #openstack-neutron-ovn | 06:15 | |
| *** markmcclain has quit IRC | 06:21 | |
| *** markmcclain has joined #openstack-neutron-ovn | 06:43 | |
| *** markmcclain has quit IRC | 06:48 | |
| *** salv-orlando has quit IRC | 06:57 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 07:09 | |
| *** markmcclain has joined #openstack-neutron-ovn | 07:17 | |
| *** markmcclain has quit IRC | 07:22 | |
| *** switchca1e is now known as switchcade | 07:33 | |
| *** salv-orlando has quit IRC | 07:43 | |
| *** yamamoto_ has quit IRC | 08:05 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 08:06 | |
| *** yamamoto_ has joined #openstack-neutron-ovn | 08:08 | |
| *** yamamoto_ has quit IRC | 08:13 | |
| *** yamamoto_ has joined #openstack-neutron-ovn | 08:23 | |
| *** yamamoto_ has quit IRC | 08:24 | |
| *** salv-orlando has quit IRC | 08:25 | |
| *** yamamoto_ has joined #openstack-neutron-ovn | 08:27 | |
| *** yamamoto_ has quit IRC | 08:28 | |
| *** yamamoto_ has joined #openstack-neutron-ovn | 08:44 | |
| *** yamamoto_ has quit IRC | 08:45 | |
| *** yamamoto_ has joined #openstack-neutron-ovn | 08:48 | |
| *** yamamoto_ has quit IRC | 08:53 | |
| *** yamamoto_ has joined #openstack-neutron-ovn | 08:56 | |
| *** yamamoto_ has quit IRC | 09:04 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 09:07 | |
| *** yamamoto_ has joined #openstack-neutron-ovn | 09:12 | |
| *** yamamoto_ has quit IRC | 09:13 | |
| *** yamamoto_ has joined #openstack-neutron-ovn | 09:21 | |
| *** fzdarsky has joined #openstack-neutron-ovn | 09:22 | |
| *** salv-orlando has quit IRC | 09:27 | |
| *** markmcclain has joined #openstack-neutron-ovn | 09:31 | |
| *** ajo has joined #openstack-neutron-ovn | 09:37 | |
| *** markmcclain has quit IRC | 09:40 | |
| *** ajo has quit IRC | 09:42 | |
| *** ajo has joined #openstack-neutron-ovn | 09:43 | |
| *** ajo has quit IRC | 09:48 | |
| *** ajo has joined #openstack-neutron-ovn | 09:48 | |
| *** ajo has quit IRC | 09:49 | |
| openstackgerrit | Numan Siddique proposed openstack/networking-ovn: Remove the debug logs in RowEvent.matches for unmatched conditions https://review.openstack.org/274574 | 09:50 |
|---|---|---|
| *** openstackgerrit has quit IRC | 10:02 | |
| *** openstackgerrit_ has joined #openstack-neutron-ovn | 10:02 | |
| *** openstackgerrit_ has quit IRC | 10:03 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 10:03 | |
| *** yamamoto_ has quit IRC | 10:08 | |
| *** salv-orlando has quit IRC | 10:40 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 10:50 | |
| *** yamamoto has joined #openstack-neutron-ovn | 11:17 | |
| *** yamamoto has quit IRC | 11:23 | |
| *** yamamoto has joined #openstack-neutron-ovn | 11:24 | |
| *** yamamoto has quit IRC | 11:28 | |
| *** yamamoto has joined #openstack-neutron-ovn | 11:31 | |
| *** yamamoto has quit IRC | 11:36 | |
| *** openstackgerrit has joined #openstack-neutron-ovn | 11:53 | |
| *** openstackgerrit has quit IRC | 11:54 | |
| *** openstackgerrit_ has joined #openstack-neutron-ovn | 11:54 | |
| *** openstackgerrit_ is now known as openstackgerrit | 11:55 | |
| *** openstackgerrit has quit IRC | 11:59 | |
| *** salv-orl_ has joined #openstack-neutron-ovn | 12:03 | |
| *** salv-orlando has quit IRC | 12:06 | |
| *** openstackgerrit has joined #openstack-neutron-ovn | 12:07 | |
| *** rtheis has joined #openstack-neutron-ovn | 12:27 | |
| *** salv-orl_ has quit IRC | 12:47 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 13:02 | |
| Sam-I-Am | mornings | 13:40 |
| openstackgerrit | Matthew Kassawara proposed openstack/networking-ovn: doc: Provide a method for access to private net. https://review.openstack.org/274236 | 13:45 |
| *** yamamoto has joined #openstack-neutron-ovn | 13:46 | |
| *** yamamoto has quit IRC | 13:55 | |
| *** yamamoto has joined #openstack-neutron-ovn | 14:03 | |
| *** jckasper has joined #openstack-neutron-ovn | 14:03 | |
| *** yamamoto has quit IRC | 14:05 | |
| *** yamamoto has joined #openstack-neutron-ovn | 14:08 | |
| openstackgerrit | Merged openstack/networking-ovn: Fix format of externals whitelist https://review.openstack.org/274315 | 14:29 |
| openstackgerrit | Merged openstack/networking-ovn: Vagrant: Consistency with upstream docs https://review.openstack.org/274254 | 14:29 |
| *** manand has joined #openstack-neutron-ovn | 14:51 | |
| *** yamamoto has quit IRC | 14:54 | |
| *** ajo has joined #openstack-neutron-ovn | 15:01 | |
| *** regXboi has joined #openstack-neutron-ovn | 15:05 | |
| * regXboi wanders in | 15:06 | |
| Mic22 | Hi russellb: how should I debug a situation where | 15:10 |
| Mic22 | i can ping dhcp agent from host but the instance cannot reach dhcp | 15:10 |
| Mic22 | and with static conf cannot reach anything else? | 15:11 |
| russellb | Mic22: what are your security groups set to? | 15:11 |
| Sam-I-Am | Mic22: where is the vm running? | 15:11 |
| Mic22 | allow everyting TCP UDP ICMP | 15:12 |
| Mic22 | on the same host | 15:12 |
| Mic22 | i'm trying to not use devstack but configure all by source | 15:12 |
| Sam-I-Am | Mic22: is this a provider network? | 15:14 |
| Mic22 | no | 15:14 |
| Mic22 | i see the ports correctly created in ovs and the addresses appear properly in the northd | 15:15 |
| russellb | so single node setup, VM and dhcp agent on the same host | 15:15 |
| Mic22 | exacly | 15:15 |
| russellb | and you've created a port on your host to let you connect in to that network | 15:15 |
| Mic22 | yep | 15:16 |
| Mic22 | from that port i can reach dhcp | 15:16 |
| Mic22 | as always if you want ssh to see by yourself pvt | 15:17 |
| *** yamamoto has joined #openstack-neutron-ovn | 15:20 | |
| Mic22 | incpecting with tshark i see from the instance with static config the ICMP to the destination but reply is not generated | 15:22 |
| *** numans has quit IRC | 15:25 | |
| *** numans has joined #openstack-neutron-ovn | 15:26 | |
| openstackgerrit | Russell Bryant proposed openstack/networking-ovn: Explicitly delete DHCP ports in delete_network. https://review.openstack.org/274732 | 15:54 |
| * russellb posts that patch out of desperation ... | 15:55 | |
| Sam-I-Am | russellb: errrr? | 15:56 |
| *** salv-orlando has quit IRC | 15:57 | |
| mestery | russellb: Looking ... | 15:58 |
| openstackgerrit | Russell Bryant proposed openstack/networking-ovn: Explicitly delete DHCP ports in delete_network. https://review.openstack.org/274732 | 15:58 |
| russellb | just a typo fix in v2 | 15:58 |
| * mestery ponders that patch | 15:59 | |
| mestery | russellb: Thanks for digging into this! | 15:59 |
| * regXboi isn't sure he's reading that patch correctly | 15:59 | |
| * Sam-I-Am ponders too | 15:59 | |
| russellb | it's also possible that i'm just making delete_network() take longer and the error less likely to hit | 15:59 |
| mestery | russellb: Do you think this is a bandage (that also happens to be bandaged up in ML2)? | 15:59 |
| russellb | yes | 15:59 |
| russellb | i do | 15:59 |
| mestery | russellb: Right | 15:59 |
| mestery | OK | 15:59 |
| russellb | i could try making the change in the base plugin and see what happens | 16:00 |
| regXboi | I almost think that would be better | 16:00 |
| russellb | changing base plugin? | 16:00 |
| regXboi | multiple bandages scare me | 16:00 |
| Sam-I-Am | fewer workarounds = mo betta | 16:00 |
| russellb | yeah.. | 16:00 |
| Sam-I-Am | wouldnt surprise me if something is fundamentally broken somewhere | 16:00 |
| russellb | the ML2 plugin has some cleanup while loops | 16:01 |
| regXboi | otoh, I'd say let's push this through in parallel, just in case the base plugin patch starts to bike shed | 16:01 |
| russellb | that i think are just working around weird stuff | 16:01 |
| *** markmcclain has joined #openstack-neutron-ovn | 16:01 | |
| *** numans has quit IRC | 16:01 | |
| russellb | maybe i should at a #TODO this is a evil hack of a workaround that may not even be necessary | 16:01 |
| regXboi | :) | 16:02 |
| russellb | I'm tired of seeing NetworkInUse failures! | 16:04 |
| russellb | must squash them! | 16:04 |
| openstackgerrit | Russell Bryant proposed openstack/networking-ovn: Explicitly delete DHCP ports in delete_network. https://review.openstack.org/274732 | 16:05 |
| russellb | ok todo added ... | 16:05 |
| russellb | i'll patch the base plugin in parallel to see what happens | 16:06 |
| *** shettyg has joined #openstack-neutron-ovn | 16:13 | |
| *** thumpba has joined #openstack-neutron-ovn | 16:19 | |
| *** roeyc has joined #openstack-neutron-ovn | 16:22 | |
| *** tongli has joined #openstack-neutron-ovn | 16:22 | |
| *** roeyc has quit IRC | 16:26 | |
| tongli | hi, guys, I run vagrant script and setup 4 nodes ovn devstack. spin up couple of VMs. each VM will get private IP. | 16:27 |
| openstackgerrit | Russell Bryant proposed openstack/networking-ovn: Drop auto-delete DHCP port workaround. https://review.openstack.org/274768 | 16:27 |
| tongli | then I associate floating IP to the vm as well, but can not ping any of the VM using floating IP. Wonder if something else has to be done to be able to access the VM using floating IP. | 16:27 |
| Sam-I-Am | russellb: off the top of your head, how much overhead does geneve consume with ovn? | 16:27 |
| Sam-I-Am | tongli: that stuff is broken | 16:28 |
| russellb | Sam-I-Am: i calculated it once and i don't remember | 16:28 |
| Sam-I-Am | native l3 support does not work yet | 16:28 |
| russellb | Sam-I-Am: 80-ish bytes? | 16:28 |
| russellb | Sam-I-Am: for testing, call it 100 and you'll be safe | 16:28 |
| Sam-I-Am | russellb: ok. i'm trying to calculate mtu for vms | 16:28 |
| tongli | @Sam-I-Am, because the flow table entries were not setup right? | 16:28 |
| russellb | tongli: the vagrant setup uses OVN native L3, which does not yet support NAT or floating IPs | 16:29 |
| russellb | so it's just not going to work ... | 16:29 |
| russellb | tongli: you can get access to the private network your VMs are on using another hack though | 16:29 |
| Sam-I-Am | routing without nat doesnt work either | 16:29 |
| tongli | @russellb,@Sam-I-Am, if I am trying to make it work, I will need to use the neutron l3 agent? | 16:30 |
| russellb | tongli: yes | 16:30 |
| Sam-I-Am | tongli: yes, but that depends on what you're trying to test | 16:30 |
| russellb | Sam-I-Am: it should, just not from provider network | 16:30 |
| russellb | within OVN networks it does | 16:30 |
| Sam-I-Am | russellb: right | 16:30 |
| tongli | @russellb, @Sam-I-Am, is there a blueprint or some work item to make it working with OVN L3 mode? | 16:32 |
| tongli | if it is not all difficult, I would like to look at it and try to work on. | 16:32 |
| russellb | we have a bug open for floating IP support | 16:32 |
| russellb | it's difficult :) | 16:32 |
| Sam-I-Am | tongli: a lot of native l3 things need work | 16:33 |
| russellb | but different pieces are being actively worked on by at least 3 people | 16:33 |
| russellb | 4 really | 16:33 |
| tongli | @russellb, cool. cool. Thanks. so there have been some draft patches exist? | 16:33 |
| russellb | for floating IPs, yes | 16:34 |
| russellb | though not sure they have been shared | 16:34 |
| russellb | there's a design patch that just got posted | 16:34 |
| russellb | but i haven't even had time to open it uyet | 16:34 |
| tongli | @russellb, thanks, I will look for them. great day! | 16:35 |
| regXboi | russellb: yeah, that's on my list to review today | 16:37 |
| * regXboi waits for computer updates to finish | 16:37 | |
| * Sam-I-Am tries the russellb private net hack | 16:38 | |
| russellb | there's no amount of caffeine that can save me today | 16:38 |
| russellb | Sam-I-Am: heh | 16:38 |
| russellb | Sam-I-Am: i actually got the idea from Mic22 :) | 16:38 |
| Sam-I-Am | ack, there's a bug :/ | 16:41 |
| *** salv-orlando has joined #openstack-neutron-ovn | 16:42 | |
| regXboi | no | 16:43 |
| regXboi | wrong window :) | 16:43 |
| russellb | Sam-I-Am: what's the bug | 16:50 |
| Sam-I-Am | the ip for 'ip addr add' lacks a prefix, so it assumes /32. | 16:50 |
| Sam-I-Am | i'm fixerating | 16:51 |
| russellb | ah | 16:51 |
| Sam-I-Am | should have caught that during my first review, but oh well | 16:52 |
| Sam-I-Am | actually dont need the route add either | 16:52 |
| russellb | Sam-I-Am: i probably had to add the route because of that | 16:56 |
| * russellb is thankful Sam-I-Am is around to clean up after him | 16:56 | |
| russellb | :-/ ;-) | 16:56 |
| Sam-I-Am | russellb: route should have returned an error | 16:57 |
| russellb | didn't for me | 16:57 |
| russellb | who knows what i was doing then | 16:57 |
| russellb | something worked! | 16:58 |
| russellb | i then tried to write down what i did | 16:58 |
| openstackgerrit | Matthew Kassawara proposed openstack/networking-ovn: doc: Provide a method for access to private net. https://review.openstack.org/274236 | 16:58 |
| russellb | maybe i didn't translate properly | 16:58 |
| russellb | ¯\_(ツ)_/¯ | 16:58 |
| Sam-I-Am | now it works :) | 16:58 |
| *** yamamoto has quit IRC | 16:58 | |
| russellb | Mic22: details in PM, but the problem is that you're not using an openvswitch kernel module with OVS+conntrack support | 16:59 |
| Sam-I-Am | russellb: so, good news... with that hack and not-yet-uploaded mods to the vagrant bits, i can see the vm | 17:00 |
| russellb | huzzah | 17:00 |
| * russellb waits for bad news | 17:01 | |
| Sam-I-Am | my vagrant mods move metadata and dhcp to the compute nodes, and also run > 1 | 17:01 |
| Sam-I-Am | which is a more realistic test | 17:01 |
| Sam-I-Am | removed n-cpu, dhcp, and md from the controller | 17:01 |
| russellb | how much RAM does it take to run this vagrant setup | 17:02 |
| Sam-I-Am | russellb: roughly 10 gigs by defaut | 17:03 |
| russellb | ok | 17:03 |
| Sam-I-Am | ooo, and my instance on the provider network can reach the interwebs | 17:03 |
| Sam-I-Am | and i can get to it from the host | 17:04 |
| russellb | bbl, LUNCH | 17:04 |
| *** armax has joined #openstack-neutron-ovn | 17:05 | |
| tongli | @russellb, it normally take me about 1 hour to get the four nodes ovn devstack setup. I have 16GB ram macbook pro. | 17:08 |
| Sam-I-Am | tongli: it takes me about 1.5 hours on a linux box with 64 gigs. the biggest variable so far is how long devstack repo cloning takes | 17:09 |
| Sam-I-Am | and pulling down packages | 17:09 |
| tongli | @russellb, once the setup is done, the machine respond well, spin up VM is very fast, I feel it a lot fast. | 17:09 |
| *** dlundquist has joined #openstack-neutron-ovn | 17:09 | |
| Sam-I-Am | tongli: i'm about to submit another patch to make things faster and more realistic | 17:09 |
| tongli | @Sam-I-Am, it seems to me that the vagrant script take a long time to build openvswitch stuff from the source. | 17:10 |
| Sam-I-Am | yeah thats pretty slow too, but consistent | 17:10 |
| tongli | and it will do that for all 4 nodes, so that added a lot of time. | 17:10 |
| Sam-I-Am | it would be nice to build it once | 17:11 |
| *** dlundquist1 has joined #openstack-neutron-ovn | 17:11 | |
| tongli | hmmm, I can add something there for it if you guys are ok with it. | 17:11 |
| Sam-I-Am | tongli: sure. i think the vagrant shared directory might be useful here. | 17:12 |
| openstackgerrit | Merged openstack/networking-ovn: Remove the debug logs in RowEvent.matches for unmatched conditions https://review.openstack.org/274574 | 17:13 |
| *** dlundquist has quit IRC | 17:14 | |
| tongli | sure. let me look at that see if we want to use a prebuilt ovs stuff or build on spot. | 17:15 |
| *** dlundquist1 is now known as dlundquist | 17:15 | |
| Sam-I-Am | we need to build ovs/ovn from source at least once | 17:16 |
| tongli | what I have when I do the test, is build ovs/ovn from source and save all the deb files. | 17:16 |
| openstackgerrit | Matthew Kassawara proposed openstack/networking-ovn: Vagrant: Modify OpenStack services https://review.openstack.org/274795 | 17:17 |
| tongli | then I use these deb files 2.5.90 for later build, I do not build over and over | 17:17 |
| tongli | for later install. | 17:17 |
| tongli | if we have to have build it per install/setup, we can build it on ovsdb node which is the first node to setup. | 17:18 |
| Sam-I-Am | yeah, that makes sense | 17:20 |
| Sam-I-Am | probably just build it into the shared dir and let the other nodes see it | 17:22 |
| *** allan_h has joined #openstack-neutron-ovn | 17:33 | |
| *** salv-orl_ has joined #openstack-neutron-ovn | 17:36 | |
| *** salv-orlando has quit IRC | 17:39 | |
| *** numans has joined #openstack-neutron-ovn | 17:48 | |
| *** chandrav has joined #openstack-neutron-ovn | 17:55 | |
| *** arosen has joined #openstack-neutron-ovn | 17:57 | |
| regXboi | russellb: ping? | 18:00 |
| *** dlundquist has quit IRC | 18:07 | |
| *** dlundquist has joined #openstack-neutron-ovn | 18:11 | |
| openstackgerrit | Amitabha Biswas proposed openstack/networking-ovn: Floating IP support design (ready for review). https://review.openstack.org/274274 | 18:19 |
| *** azbiswas has joined #openstack-neutron-ovn | 18:20 | |
| russellb | regXboi: pong | 18:33 |
| regXboi | russellb: have you had any time to look at bug 1536003 since the last update? | 18:39 |
| openstack | bug 1536003 in networking-ovn "high cpu usage on ovn-controller process" [High,Confirmed] https://launchpad.net/bugs/1536003 - Assigned to Russell Bryant (russellb) | 18:39 |
| russellb | no | 18:40 |
| russellb | i had one patch that helped ovn-controller a bit | 18:40 |
| russellb | the major performance improvements for ovn-northd and ovn-controller are actually pretty well known, though | 18:40 |
| russellb | it sounds like you're able to break networking-ovn though, we need to figure that out | 18:41 |
| russellb | but i haven't looked in the last week | 18:41 |
| Mic22 | guys if I would like to have ovn-northd highly available do I have to put the database files on a shared drive and make an active/passive or just make an active/passive | 18:41 |
| Mic22 | ? | 18:41 |
| regXboi | russellb: ok, I'll powwow with mestery this afternoon on it | 18:43 |
| russellb | regXboi: ok | 18:43 |
| *** s3wong has joined #openstack-neutron-ovn | 18:45 | |
| openstackgerrit | Russell Bryant proposed openstack/networking-ovn: Add more thoughts to the HA section of the FAQ. https://review.openstack.org/274844 | 18:48 |
| russellb | Mic22: short answer, yes | 18:48 |
| Mic22 | good | 18:49 |
| russellb | Mic22: long answer, see the FAQ, including that patch which adds a bit more commentary to the HA section of the FAQ | 18:49 |
| Mic22 | mmm i read the faq some times ago is that new? | 18:49 |
| russellb | not really | 18:49 |
| russellb | added early November | 18:49 |
| Mic22 | ok sorry my brain must be fused | 18:51 |
| *** dlundquist has quit IRC | 18:52 | |
| russellb | blame your browser cache :) | 18:52 |
| *** dlundquist has joined #openstack-neutron-ovn | 18:54 | |
| russellb | regXboi: fixed your ovn-northd logging bug :) https://bugs.launchpad.net/networking-ovn/+bug/1539994 | 18:54 |
| openstack | Launchpad bug 1539994 in networking-ovn "ovn-northd does not log to console" [Low,Fix released] - Assigned to Russell Bryant (russellb) | 18:54 |
| russellb | low hanging fruit fixes feel good | 18:54 |
| Sam-I-Am | instant gratification | 18:54 |
| russellb | nice to have those mixed in with other much more frustrating things ... | 18:54 |
| regXboi | russellb: yay! I'll restack and try it out | 18:54 |
| russellb | great, thanks | 18:54 |
| regXboi | russellb: and then update the devstack plugin accordingly :) | 18:55 |
| russellb | cool | 18:55 |
| russellb | it actually should magically start working | 18:55 |
| russellb | i think | 18:55 |
| russellb | yes | 18:55 |
| russellb | it will now log to the console by default | 18:55 |
| openstackgerrit | Numan Siddique proposed openstack/networking-ovn: Sync neutron db and OVN NB db from the Ovn Worker context https://review.openstack.org/274849 | 18:57 |
| regXboi | cool! | 18:57 |
| *** salv-orl_ has quit IRC | 19:00 | |
| regXboi | russellb: ping - got a conntrack question for you this time :) | 19:03 |
| regXboi | actually conntrack/GENEVE :) | 19:03 |
| russellb | ask away | 19:03 |
| russellb | i may or may not answer | 19:03 |
| regXboi | do you know of a patch that would allow conntrack to look inside the GENEVE packet when matching connections | 19:04 |
| regXboi | I'm trying to avoid the situation where | 19:04 |
| regXboi | let's say I wrap a TCP request in GENEVE -> I get a UDP packet going from host A/port X to host B/port 6081 | 19:04 |
| regXboi | 1:00 | 19:04 |
| regXboi | the TCP response will also be wrapped in GENEVE and I'll get a UDP packet going from host B/port Y to host A/port 6081 | 19:04 |
| regXboi | and now I have two Unreplied entries in conntrack | 19:04 |
| *** dlundquist has quit IRC | 19:04 | |
| russellb | where exactly are these packets getting tracked | 19:04 |
| russellb | that shouldn't be happening with how OVN uses conntrack. | 19:05 |
| *** dlundquist has joined #openstack-neutron-ovn | 19:05 | |
| regXboi | ok... | 19:05 |
| russellb | i know i didn't answer your question | 19:06 |
| russellb | :) | 19:06 |
| regXboi | no, actually, you did | 19:06 |
| russellb | ok | 19:06 |
| regXboi | sorta | 19:06 |
| russellb | sorta | 19:06 |
| regXboi | or at least enough to chase the original source back to getting me more information | 19:06 |
| russellb | ok, right | 19:06 |
| Sam-I-Am | russellb: it should be tracking pre-tunnel, no? | 19:07 |
| russellb | right | 19:07 |
| Sam-I-Am | ok, thats what i was thinking | 19:07 |
| russellb | basically those entries are outside of what OVN does | 19:07 |
| regXboi | yeah, I've gone back to the original source and asked where these entries are being seen | 19:08 |
| regXboi | because I suspect the folks seeing them are looking in the wrong place | 19:08 |
| openstackgerrit | Matthew Kassawara proposed openstack/networking-ovn: Vagrant: Add docs for instance external net access https://review.openstack.org/274856 | 19:09 |
| *** gangil has joined #openstack-neutron-ovn | 19:09 | |
| *** gangil has joined #openstack-neutron-ovn | 19:09 | |
| *** chandrav has quit IRC | 19:16 | |
| *** gangil has quit IRC | 19:20 | |
| *** gangil has joined #openstack-neutron-ovn | 19:23 | |
| *** gangil has joined #openstack-neutron-ovn | 19:23 | |
| *** thumpba has quit IRC | 19:23 | |
| *** thumpba has joined #openstack-neutron-ovn | 19:26 | |
| openstackgerrit | Numan Siddique proposed openstack/networking-ovn: Create dhcp agent scheduler and periodic status check in Ovn Worker https://review.openstack.org/274864 | 19:31 |
| *** numans has quit IRC | 19:40 | |
| openstackgerrit | Russell Bryant proposed openstack/networking-ovn: tempest debugging, ignore... https://review.openstack.org/269705 | 19:43 |
| *** gongysh has joined #openstack-neutron-ovn | 19:44 | |
| openstackgerrit | Russell Bryant proposed openstack/networking-ovn: tempest debugging, ignore... https://review.openstack.org/269705 | 19:49 |
| *** azbiswas has quit IRC | 19:55 | |
| *** azbiswas has joined #openstack-neutron-ovn | 19:56 | |
| *** roeyc has joined #openstack-neutron-ovn | 19:57 | |
| *** azbiswas has quit IRC | 20:00 | |
| *** jckasper has quit IRC | 20:02 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 20:03 | |
| *** salv-orlando has quit IRC | 20:07 | |
| regXboi | russellb: ack on the northd logging fix - now adding the code to change to include msec timestamps on console :) | 20:09 |
| russellb | cool | 20:09 |
| russellb | thanks for testing | 20:09 |
| openstackgerrit | Ryan Moats proposed openstack/networking-ovn: Add msec resolution to ovn-northd console logs. https://review.openstack.org/274876 | 20:15 |
| regXboi | and there we go | 20:15 |
| russellb | tested? | 20:22 |
| *** tongli has quit IRC | 20:28 | |
| regXboi | yes, I tested it on an AIO node before I pushed it up | 20:32 |
| *** markmcclain has quit IRC | 20:32 | |
| regXboi | but let's make sure I didn't do something stupid with red thingies that jenkins objects to | 20:32 |
| *** markmcclain has joined #openstack-neutron-ovn | 20:33 | |
| *** chandrav has joined #openstack-neutron-ovn | 20:37 | |
| *** azbiswas has joined #openstack-neutron-ovn | 20:38 | |
| *** gongysh has quit IRC | 20:38 | |
| *** salv-orlando has joined #openstack-neutron-ovn | 20:53 | |
| *** jckasper has joined #openstack-neutron-ovn | 20:55 | |
| *** jckasper_ has joined #openstack-neutron-ovn | 20:55 | |
| *** jckasper_ has quit IRC | 20:59 | |
| regXboi | o/ | 20:59 |
| *** jckasper has quit IRC | 20:59 | |
| regXboi | \o | 20:59 |
| *** jckasper has joined #openstack-neutron-ovn | 20:59 | |
| *** thumpba has quit IRC | 21:00 | |
| * regXboi practices semaphore | 21:00 | |
| *** fzdarsky is now known as fzdarsky|afk | 21:07 | |
| Sam-I-Am | russellb: would there be any case where private networks wouldnt use an overlay? | 21:09 |
| *** markmcclain has quit IRC | 21:09 | |
| *** markmcclain has joined #openstack-neutron-ovn | 21:09 | |
| russellb | Sam-I-Am: well ... depends what your definition of private is? | 21:10 |
| russellb | provider network API extension is admin-only by default | 21:11 |
| Sam-I-Am | russellb: not provider | 21:11 |
| russellb | so if you mean tenant-created self-service networks, they would always be an overlay in OVN land | 21:11 |
| Sam-I-Am | ok thats what i wanted to know. in conventional neutron, you can have vlan and flat tenant networks | 21:11 |
| russellb | ah yes, that's right | 21:12 |
| russellb | OVN does not support that | 21:12 |
| russellb | if we did, we'd have reason to use ML2 ... | 21:12 |
| russellb | but alas | 21:12 |
| Sam-I-Am | looks like geneve overhead in ovn is 42 bytes | 21:13 |
| *** markmcclain has quit IRC | 21:27 | |
| regXboi | russellb: in re https://patchwork.ozlabs.org/patch/571318/ | 21:33 |
| russellb | yar | 21:34 |
| regXboi | does that depends on another patch to compile correctly? | 21:34 |
| regXboi | er *depend* | 21:34 |
| russellb | it might.. | 21:34 |
| regXboi | the change to ovn-controller.c doesn't go straight into master because local_datapaths isn't defined anywhere :/ | 21:34 |
| russellb | ah. | 21:34 |
| russellb | https://github.com/russellb/ovs/commits/ovn-controller-perf | 21:34 |
| russellb | i wrote it on top of localnet port fixes | 21:35 |
| russellb | since your env was using those patches already | 21:35 |
| regXboi | aha! | 21:35 |
| *** markmcclain has joined #openstack-neutron-ovn | 21:35 | |
| regXboi | ok, me rolls forward another patch set | 21:35 |
| *** arosen has quit IRC | 21:41 | |
| *** arosen1 has joined #openstack-neutron-ovn | 21:41 | |
| openstackgerrit | Merged openstack/networking-ovn: Sync neutron db and OVN NB db from the Ovn Worker context https://review.openstack.org/274849 | 21:42 |
| *** arosen has joined #openstack-neutron-ovn | 21:45 | |
| *** arosen is now known as Guest65685 | 21:45 | |
| *** arosen1 has quit IRC | 21:48 | |
| *** Guest65685 has quit IRC | 21:48 | |
| *** Guest65685 has joined #openstack-neutron-ovn | 21:48 | |
| *** Guest65685 has quit IRC | 21:48 | |
| *** thumpba has joined #openstack-neutron-ovn | 22:02 | |
| *** dlundquist has quit IRC | 22:25 | |
| *** jckasper has quit IRC | 22:46 | |
| *** shettyg has quit IRC | 22:59 | |
| *** regXboi has quit IRC | 23:05 | |
| *** jckasper has joined #openstack-neutron-ovn | 23:09 | |
| *** jckasper has quit IRC | 23:14 | |
| *** armax has quit IRC | 23:15 | |
| *** azbiswas has quit IRC | 23:19 | |
| *** azbiswas has joined #openstack-neutron-ovn | 23:19 | |
| *** gangil has quit IRC | 23:22 | |
| *** azbiswas has quit IRC | 23:24 | |
| *** armax has joined #openstack-neutron-ovn | 23:29 | |
| *** roeyc has quit IRC | 23:30 | |
| *** gangil has joined #openstack-neutron-ovn | 23:36 | |
| *** gangil has joined #openstack-neutron-ovn | 23:36 | |
| *** salv-orl_ has joined #openstack-neutron-ovn | 23:37 | |
| *** dlundquist has joined #openstack-neutron-ovn | 23:39 | |
| *** salv-orlando has quit IRC | 23:40 | |
| *** dlundquist has quit IRC | 23:42 | |
| *** dlundquist has joined #openstack-neutron-ovn | 23:45 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!