Thursday, 2016-02-04

*** azbiswas has joined #openstack-neutron-ovn00:23
*** chenli has joined #openstack-neutron-ovn01:03
*** chandrav has joined #openstack-neutron-ovn01:19
*** azbiswas has quit IRC01:47
*** yamamoto_ has joined #openstack-neutron-ovn01:48
*** chenli_ has joined #openstack-neutron-ovn01:54
*** chenli has quit IRC01:57
*** chenli_ is now known as chenli01:57
*** yamamoto_ has quit IRC01:59
*** yamamoto_ has joined #openstack-neutron-ovn02:00
*** yamamoto_ has quit IRC02:10
*** yamamoto_ has joined #openstack-neutron-ovn02:10
*** yamamoto_ has quit IRC02:13
*** s3wong has quit IRC02:40
*** chandrav has quit IRC02:46
*** jckasper has quit IRC03:02
*** fzdarsky|afk has quit IRC03:02
*** jckasper has joined #openstack-neutron-ovn03:03
*** jckasper has quit IRC03:03
*** jckasper has joined #openstack-neutron-ovn03:03
*** fzdarsky has joined #openstack-neutron-ovn03:05
*** yamamoto_ has joined #openstack-neutron-ovn03:09
*** allan_h has quit IRC03:16
*** azbiswas has joined #openstack-neutron-ovn03:31
*** thumpba has joined #openstack-neutron-ovn03:34
mesterySam-I-Am: This tox genconfig mess is really starting to irritate me :(03:41
*** thumpba has quit IRC03:53
*** thumpba has joined #openstack-neutron-ovn03:54
*** salv-orl_ has joined #openstack-neutron-ovn03:54
*** salv-orlando has quit IRC03:58
*** azbiswas has quit IRC03:58
*** thumpba_ has joined #openstack-neutron-ovn04:12
*** thumpba has quit IRC04:13
*** russellb has quit IRC04:19
*** russellb has joined #openstack-neutron-ovn04:20
*** thumpba_ has quit IRC04:25
Sam-I-Ammestery: my patch passed the gate last i looked04:27
*** azbiswas has joined #openstack-neutron-ovn04:59
*** azbiswas has quit IRC05:05
*** azbiswas has joined #openstack-neutron-ovn05:10
openstackgerritMatthew Kassawara proposed openstack/networking-ovn: Testing tox
*** jckasper has quit IRC05:31
*** jckasper has joined #openstack-neutron-ovn05:32
*** jckasper has quit IRC05:50
*** jckasper has joined #openstack-neutron-ovn05:51
*** jckasper has quit IRC05:56
*** numans has joined #openstack-neutron-ovn06:02
*** armax has quit IRC06:20
openstackgerritlitong01 proposed openstack/networking-ovn: Vagrant: Increase the size of the provider network subnet
*** tongli has quit IRC06:30
*** azbiswas has quit IRC06:45
*** salv-orl_ has quit IRC07:13
*** gangil has quit IRC07:17
*** roeyc has joined #openstack-neutron-ovn07:55
*** fzdarsky has quit IRC08:38
*** fzdarsky has joined #openstack-neutron-ovn08:38
*** openstackgerrit has quit IRC09:17
*** openstackgerrit has joined #openstack-neutron-ovn09:17
*** numans has quit IRC09:23
*** chenli has quit IRC09:36
*** jckasper has joined #openstack-neutron-ovn09:52
*** jckasper has quit IRC09:57
Mic22russellb: I finally found out what the error was! If you put a wrong UUID as the iface-id and than you put the right ovn-controller take minutes to properly update the flows10:01
Mic22practically the first flows are generated an all the computes for a wrong port and then when you put the write uuid the local flows are configured to drop the packet10:02
openstackgerritBabu Shanmugam proposed openstack/networking-ovn: DPDK support for OVN
*** palexster has quit IRC10:32
*** salv-orlando has joined #openstack-neutron-ovn10:54
*** roeyc has quit IRC11:08
*** yamamoto_ has quit IRC11:29
*** roeyc has joined #openstack-neutron-ovn11:44
*** chandrav has joined #openstack-neutron-ovn12:04
*** rtheis has joined #openstack-neutron-ovn12:12
*** numans has joined #openstack-neutron-ovn12:24
*** palexster has joined #openstack-neutron-ovn12:28
*** salv-orlando has quit IRC12:32
*** roeyc has quit IRC12:45
*** salv-orlando has joined #openstack-neutron-ovn12:46
*** numan_ has joined #openstack-neutron-ovn12:52
*** numans has quit IRC12:53
*** roeyc has joined #openstack-neutron-ovn12:55
*** jckasper has joined #openstack-neutron-ovn12:57
*** jckasper has quit IRC13:09
*** jckasper has joined #openstack-neutron-ovn13:10
*** jckasper has quit IRC13:14
*** yamamoto has joined #openstack-neutron-ovn13:31
russellbMic22: I'm glad you got it sorted out :)13:34
*** yamamoto has quit IRC13:43
numan_russellb, Hi, can you please review this again when you get some time -
russellbnuman_: done13:46
russellbneed someone else to approve13:46
numan_russellb, thank you13:46
* russellb looks around13:46
* Sam-I-Am stumbles in13:46
russellbSam-I-Am: want to approve that?  you can blame me if it breaks13:47
Sam-I-Amrussellb: seems logical, pre-coffee13:50
*** jckasper has joined #openstack-neutron-ovn13:51
* Sam-I-Am wonders if +2 makes him coffee13:51
rtheisSam-I-Am: Any luck so far with the tox genconfig issue?13:52
Sam-I-Amrtheis: well, my patch from yesterday has worked a couple of times13:53
Sam-I-Amandreas has yet another suggestion13:53
*** jckasper has quit IRC13:53
Sam-I-Amstill dont really know why it broke all of the sudden13:53
*** jckasper has joined #openstack-neutron-ovn13:53
Sam-I-Amonly thing i can suspect is either a) something was updated in the image or b) we got into some nodes with slightly different images13:54
openstackgerritMatthew Kassawara proposed openstack/networking-ovn: Fix gate problem with tox
*** chandrav has quit IRC14:16
*** chandrav_ has joined #openstack-neutron-ovn14:17
*** salv-orlando has quit IRC14:20
*** chandrav_ has quit IRC14:22
*** chandrav has joined #openstack-neutron-ovn14:23
*** chandrav has quit IRC14:23
*** salv-orlando has joined #openstack-neutron-ovn14:24
openstackgerritMerged openstack/networking-ovn: doc: Provide a method for access to private net.
openstackgerritMatthew Kassawara proposed openstack/networking-ovn: Devstack: Improve DHCP and metadata agents
Sam-I-Amrussellb: ^ your comment triggered some elaboration in that patch14:30
Sam-I-Amnow to watch the latest tox patch14:31
* Sam-I-Am grabs popcorn14:31
Sam-I-Amstill not really sure what all happened there14:32
mesterySam-I-Am: Is the "tox -egenconfig" issue fixed yet?14:43
Sam-I-Ammestery: two patches on top of my tox patch passed gate twice. andreas suggested another change or so... testing that now.14:44
mesterySam-I-Am: Excellent! Links?14:44
Sam-I-Ammestery: you can try checking out my tox patch locally14:44
mesteryI'll look at them and help merge them if they are ready14:44
Sam-I-Amthe latest one hasnt gated yet, nor the things on top of it14:45
openstackgerritvenkatamahesh proposed openstack/networking-ovn: Use uppercase 'S' in word "OpenStack"
*** chandrav has joined #openstack-neutron-ovn15:04
*** roeyc has quit IRC15:05
russellbSam-I-Am: this still says "DO NOT MERGE" is that intentional?
Sam-I-Amrussellb: until it passes, i think?15:21
russellbpassed i think15:22
mesteryI +2'd it15:22
Sam-I-Amrussellb: it passed, but it also doesnt run the main gate jobs15:22
mesteryI need that it's blockign my Vagrant stuff :)15:23
*** chandrav has quit IRC15:23
Sam-I-Ammestery: did you check it out locally and try to run vagrant with it?15:23
Sam-I-Ammight be faster than the gate15:23
mesteryIt's running now15:23
Sam-I-Amoh, so it worked?15:23
mesteryI borked something15:24
mesteryI need to set something in local.conf15:24
mesteryto have it pull that review out15:24
mesteryFiguring that out now15:25
*** rtheis has quit IRC15:25
Sam-I-Ami'm testing out the addressing patch in vagrant now, so its tied up.15:26
Sam-I-Amotherwise i'd test the tox patch15:27
Sam-I-Amwe could just merge the tox patch after i change the commit message. i dont think the situation could get much worse. :)15:27
mesterySam-I-Am: OK, I've got it running now15:27
mesteryADded this to local.conf:15:27
mesteryI believe that should pull the patch into my vagrant run now15:27
Sam-I-Ami think so15:28
*** rtheis has joined #openstack-neutron-ovn15:28
Sam-I-Ami think that line is what was getting me when i was trying to test local changes to networking-ovn - forgot to change the branch from master to my local branch :/15:28
mesteryAh, that would do it15:29
Sam-I-Ami love brainos15:30
mesterySam-I-Am: I was wrong, I needed this:15:33
mesteryenable_plugin networking-ovn refs/changes/39/275939/215:33
* mestery restacks15:33
Sam-I-Ammestery: do those previous vars not actually work?15:38
mesterySam-I-Am: They do not, since this is a plugin, you specify repo and branch on the command line15:38
mesteryenable_plugin networking-ovn refs/changes/39/275939/215:38
mesteryAnd the first paste was wrong, that one works15:38
Sam-I-Amaight. sounds like something to fixerate+documenterate?15:39
Sam-I-Ammestery: fwiw, my vagrant stuff built without that tox patch, sooo... heh.15:40
Sam-I-Ami love gate mysteries.15:40
Sam-I-Ammestery: well, the addressing patch doesnt work15:42
Sam-I-Amneutron did not start on the controller15:42
Sam-I-Amneither did nova15:43
Sam-I-Amor glance15:43
Sam-I-Amoh, interesting15:43
Sam-I-Amsomehow eth0 and eth1 swapped roles15:44
Sam-I-Amand the ip did not move from ethX to br-provider15:44
Sam-I-Amoh, hmmmm15:47
mesteryI think that tong should have tacked on his changes on top of my original ones15:47
Sam-I-Amthat piece worked...15:47
mesteryMy original ones worked fine15:47
Sam-I-Amthe original ones were just hard-codedly simple15:47
Sam-I-Amthen this automation came in, and something isnt quite right15:47
Sam-I-Amthe compute nodes seem to work15:49
Sam-I-Amjust the controller is borked15:49
Sam-I-Ammestery: what were the symptoms when tox broke when you ran vagrant?15:54
mesteryStacking failed when it tried to generate the config for networking-ovn using "tox -egenconfig"15:54
Sam-I-Amdid vagrant keep stumbling along or did it die?15:55
*** shettyg has joined #openstack-neutron-ovn15:56
mesteryI hit a snag and started over :)15:59
Sam-I-Amooo networking guide meeting time16:00
openstackgerritMatthew Kassawara proposed openstack/networking-ovn: Fix gate problem with tox
Sam-I-Ammestery: ^ removed the do not merge16:12
mesterySam-I-Am: Thanks!16:13
Sam-I-Amcan i blame you if it doesnt work? :)16:13
Sam-I-Amrussellb: can you +2A 27593916:15
Sam-I-Amawesome. lets hope it works!16:16
Sam-I-Aminterestingly, the rally job seemed to be working while this was broken16:16
*** regXboi has joined #openstack-neutron-ovn16:19
*** yamamoto has joined #openstack-neutron-ovn16:23
*** allan_h has joined #openstack-neutron-ovn16:23
Sam-I-Amphew... glad this wasnt the tox thing :)16:24
mestery*sigh*, now getting keystone errors in my vagrant16:26
* mestery weeps and bangs his head on the desk16:26
Sam-I-Amhey, that piece actually worked for me16:26
Sam-I-Amwhat sort of errors?16:26
*** yamamoto has quit IRC16:27
mestery==> ovn-controller: Unable to establish connection to
mesterySam-I-Am: On 275748, I'm going to go back to my original version, and let Tong build on top16:28
mesteryI want the original merged, his optimizations can come later16:28
Sam-I-Ammestery: yeah, thats what i ended up with - neutron iself wasnt running16:28
Sam-I-Amneutron-server that is16:28
Sam-I-Amneither was glance or nova, but keystone was... via apache16:29
openstackgerritKyle Mestery proposed openstack/networking-ovn: Vagrant: Increase the size of the provider network subnet
mesterySam-I-Am: ^^^^16:30
*** salv-orlando has quit IRC16:37
Sam-I-Ammestery: and if that breaks...16:42
*** armax has joined #openstack-neutron-ovn16:46
openstackgerritRichard Theis proposed openstack/networking-ovn: Add network availability zone support
*** Guest55835 has quit IRC16:52
*** s3wong has joined #openstack-neutron-ovn17:11
Sam-I-Amrussellb: any progress on ?17:11
russellbi previously offerred a box of cookies as a bounty for that17:13
russellbi'm going ot have to up my offer17:13
Sam-I-Ami'll take a tesla17:15
russellbso will i17:16
russellbSam-I-Am: you should have asked for a tesla as a hiring bonus17:17
russellbcc/ mestery ^17:17
mesterylol :)17:17
Sam-I-Amyeah, what was i thinking17:17
russellbmissed opportunity17:17
* Sam-I-Am quits and asks to be re-hired, with a tesla17:17
russellbSam-I-Am: ping me if/when mestery won't re-hire you17:18
Sam-I-Amfunny thing is... i maybe drive 100 miles/month at most17:18
Sam-I-Amrussellb: got an offer with a tesla?17:18
russellbyes, (definition of "tesla" to be given by me, post-hiring)17:18
mesteryrussellb: Sneaky ;)17:18
* russellb buys a tiny tesla coil17:19
Sam-I-Ami drive a 17 year old car17:19
Sam-I-Amcant justify anything fun when i dont drive anywhere17:19
Sam-I-Amhowever, i will invest in bicycles17:19
Sam-I-Am... a 17 year old car with 95k miles on it, i might add17:20
Sam-I-Ami ride a bike about 5000 miles/year17:20
russellbi sit in my chair 3294872394 hours a year17:20
russellbmight be a rounding error in there17:21
Sam-I-Ami've been trying to figure out how to attach a laptop to my bike17:21
russellbi have
russellbok lunch ttyl.17:21
Sam-I-Ammestery: ^ does this count as a work-from-home office supply?17:22
Sam-I-Amrussellb: when you get back we need to talk about your comments on my dhcp patch17:22
openstackgerritRyan Moats proposed openstack/networking-ovn: Handle OVN NB and SB databases with separate ovsdb-server procs
regXboirussellb: in re the NB/SB separation thread ^^^^^^ :)17:28
openstackgerritRyan Moats proposed openstack/networking-ovn: Handle OVN NB and SB databases with separate ovsdb-server procs
mesteryregXboi: Need that one to include changes in the Vagrant config I think, otherwise I think that will fail (we set OVN_REMOTE there for example).17:32
regXboino you don't17:32
mesteryWe dont?17:32
* mestery looks17:32
mesteryyes we do :)17:32
regXboiyou do... just in setup-compute and setup-controller17:33
regXboinot :)17:33
mesterywas just going to point that out :)17:33
Sam-I-AmregXboi: should like 53 use ovn_sb_remote for consistency?17:33
regXboiSam-I-Am: I was trying to leave things alone17:35
Sam-I-Amwhy start now? :)17:36
regXboibecause I tend to avoiding having patches be that intrusive17:36
openstackgerritRyan Moats proposed openstack/networking-ovn: Handle OVN NB and SB databases with separate ovsdb-server procs
regXboimestery: ^^^^ should now have vagrant changes17:37
regXboiand Sam-I-Am, slap a -1 on it if it really matters to you17:37
mesteryregXboi: Does ovn-controller on compute nodes need access to the NB DB or only the SB DB?17:37
mesteryIn other words, is that only needed in the controller setup?17:37
regXboimestery: controller on compute nodes only talks to SB17:37
regXboibut the other is for running commands17:37
mesteryregXboi: OK17:38
Sam-I-AmregXboi: is this db split going to be a permanent thing?17:39
regXboiSam-I-Am: why not?17:39
*** salv-orlando has joined #openstack-neutron-ovn17:39
Sam-I-Amjust wondering if we needed to support both ways17:39
regXboiSam-I-Am: we know the old way (T) has performance issues, so mestery and I decided to default to the pipeline17:40
regXboiand I *frankly* don't see why we should keep the T around17:40
regXboibut I am open to persuasion17:40
*** brad_behle has joined #openstack-neutron-ovn17:41
Sam-I-Amyou're the performance guy, so if they need to be split, thats how we do it17:41
Sam-I-Ami'm just trying to keep the options consistent to avoid confusion17:41
openstackgerritRyan Moats proposed openstack/networking-ovn: Handle OVN NB and SB databases with separate ovsdb-server procs
regXboiSam-I-Am: done17:42
*** salv-orlando has quit IRC17:44
regXboidlundquist: I'm rather reticent about changing the SB socket name (and the SB pid as well) as that gets a bit more intrusive than I was comfortable with18:01
*** fzdarsky is now known as fzdarsky|afk18:03
*** s3wong has quit IRC18:04
*** azbiswas_ has joined #openstack-neutron-ovn18:06
russellbOVN meeting in #openvswitch in 10 minutes18:08
russellbor 7 minutes18:08
* Sam-I-Am ends up in circular dependency hell18:09
Sam-I-Ami think the devstack-vagrant interaction needs more docs18:09
*** Guest62709 has joined #openstack-neutron-ovn18:12
*** Guest62709 is now known as arosen218:13
dlundquistregXboi: okay, I'm fine leaving it as is too18:14
*** arosen2 has quit IRC18:17
*** arosen12 has joined #openstack-neutron-ovn18:17
*** arosen12 is now known as arosen118:18
*** chandrav has joined #openstack-neutron-ovn18:19
*** salv-orlando has joined #openstack-neutron-ovn18:24
openstackgerritMerged openstack/networking-ovn: Fix gate problem with tox
*** s3wong has joined #openstack-neutron-ovn18:27
*** Alik has joined #openstack-neutron-ovn18:27
Alikhi, has anyone tried to use OVN vs OpenStack to create topologies when a packet needs to travers 2 routers? I can't see OpenFlow rules installed properly to support this topology18:28
*** salv-orlando has quit IRC18:28
shettygAlik: If OVN routers, thgen I don't think they work18:29
shettygI think it is only one router that works18:29
Alikthanks shettyg, yeah, I can pass one router only. A network that connects to two routers - a packet goes to one "direction" only, I can't also see OpenFlow corresponding rules that do TTL-2, or more tables being added to accomodate the additional hop18:31
*** tongli has joined #openstack-neutron-ovn18:31
shettygAlik: yeah. I think it is known. It will need some work to fix.18:32
*** azbiswas has joined #openstack-neutron-ovn18:32
*** azbiswas_ has quit IRC18:35
Alikshettyg,  is the problem, to your knowledge, on the openstack side or on OVN side?18:35
shettygOVN side18:36
russellbthere won't be any TTL - 218:37
Alik<shettyg>, I see. Haven't tried with the OpenStack L3 plugin since I am experimenting with OVN, but it should work in that case, shouldn't it?18:37
russellbit'll be going through the same tables multiple times18:37
AlikI can't see resubmit to the same table either18:37
Sam-I-Amis there a use case for this?18:38
AlikI am just experimenting with OVN to mimic some DC virtual topologies, typically it's more complicated than just routing between 2 networks...18:39
russellbit's not a resubmit18:40
russellbit's via patch ports18:40
shettygThe thing that OVN does not have is ability learn routes across routers. So a subnet reachable via router R1 cannot be known by Router R218:40
AlikI was trying to create 2-tier routing, similar to the physical fabric-spine-leaf DC typical arch, where there can be 2 levels of routing18:42
AlikI was trying to install static routes on the router explicitly but still can't ping through the 2nd router. Actually if a network is connected to 2 routers, I can ping one direction only. I have not explored all the OpenFlow rules being installed, but I see routing tables18:44
Aliktables 18, 19 with matches for both routes, but no recirculation rules to the same table18:45
openstackgerritAaron Rosen proposed openstack/networking-ovn: test
AlikThanks Aaron, I will try that patch18:46
AlikRussell, as one of the main leads of OVN, as far as you know there is no such restriction in OVN and it should work, right?18:47
AlikRussell, I understand that each router will create a pair of patch ports, but from the pipeline (OF rules) how the flow will look like in this case?18:48
*** brad_behle has quit IRC18:57
*** allan_h has quit IRC18:57
*** allan_h has joined #openstack-neutron-ovn18:57
*** jckasper has quit IRC19:01
*** jckasper has joined #openstack-neutron-ovn19:02
*** jckasper has quit IRC19:02
*** jckasper_ has joined #openstack-neutron-ovn19:04
*** jckasper_ has quit IRC19:04
*** jckasper_ has joined #openstack-neutron-ovn19:05
*** Alik has quit IRC19:06
*** gsagie_ has joined #openstack-neutron-ovn19:11
*** gangil has joined #openstack-neutron-ovn19:21
*** gangil has joined #openstack-neutron-ovn19:21
regXboiarosen1: ping19:25
*** allan_h has quit IRC19:37
*** allan_h has joined #openstack-neutron-ovn19:38
openstackgerritKyle Mestery proposed openstack/networking-ovn: Vagrant: Workaround shebang insanities
Sam-I-Ammestery: lulz19:40
mesterySam-I-Am: I know, right?19:40
mesteryDamn Ubuntu ... Damn Docker!19:41
Sam-I-Ami havent thought about that in years19:41
mesteryTie them together and I melt19:41
Sam-I-Ambut it bit me once or twice too19:41
mesteryrussellb will appreciate my comment in 27641419:41
openstackgerritRyan Moats proposed openstack/networking-ovn: Handle OVN NB and SB databases with separate ovsdb-server procs
mesterySam-I-Am: We're heading towards the most intricate, beautiful, and surreal Vagrant config in OpenStack19:42
mesteryWatch out19:42
Sam-I-Amyeah it is becoming rather... sexy19:42
Sam-I-Amright now i'm stuck in mtu hell19:42
mesterySam-I-Am: gsagie just +2'd 276414 :)19:43
mesteryJust a hint in case you want to nudge it in19:43
russellbmestery: clearly the problem is that you're using ubuntu19:44
russellbmestery: and should switch the distro used in the vagrant setup19:44
mesteryrussellb: I know, right?19:44
regXboimestery: you so should have said something about "fixing the whole shebang" in your commit message19:44
mesteryI mean ... dash? WTF?19:44
mesteryregXboi: slow clap19:44
russellbi'd work on ovn, but i'm busy bike shedding about paperwork19:45
regXboimestery: I was referring to the song :)19:45
Sam-I-Am+2A'd wait on jenkins19:45
mesteryheh :)19:45
Sam-I-Amdont blame me if it breaks, i'm the core-fng19:46
regXboiSam-I-Am: we're kicking things up a notch?19:46
* regXboi approaches the point of failures in bug 154200519:47
openstackbug 1542005 in networking-ovn "adding compute nodes impacts ovn plugin performance under scale" [Undecided,New]
regXboiand it looks like the link update launchpad is kablooey19:48
*** allan_h has quit IRC19:48
*** salv-orlando has joined #openstack-neutron-ovn19:54
arosen1regXboi: pong19:56
regXboiarosen1: replies to your comments on 276364 attached ...19:56
arosen1regXboi:  where are there two processes launched?19:57
*** allan_h has joined #openstack-neutron-ovn19:57
Sam-I-Ammestery: vagrant seems to work with your static-larger-network patch19:58
mesterySam-I-Am: Lets get that merged then :)19:58
regXboiarosen1: where as in which node or where as in where in the code?19:58
arosen1regXboi:  does passing in multiple --remote=ptcp:PORT:HOST cause ovsdb-server to fork off additional processes to service each port?19:58
regXboiarosen1: ah ... no19:58
regXboithere are two ovsdb-server calls19:58
Sam-I-Ammestery: done!19:59
Sam-I-Amnow i'm out for a bit19:59
mesterySam-I-Am: Have fun, and thanks!19:59
regXboiarosen1: Ls 264 and 28419:59
regXboiarosen1: L284 is the original which used to do both NB and SB, L264 is the new one for NB20:00
arosen1ah okay i see sorry20:00
regXboiarosen1: no worries20:00
regXboiI'm going to need to update the commit message anyway as it looks like this closes the bug, rather than partials the bug20:00
arosen1one sec i got some more feedback. I see what i was confused.20:02
arosen1i'll just follow up on there review.20:02
regXboiarosen1: fine - like I said, PS5 isn't the end20:02
openstackgerritMerged openstack/networking-ovn: Vagrant: Increase the size of the provider network subnet
arosen1regXboi: replied20:06
*** jckasper_ has quit IRC20:27
openstackgerritRyan Moats proposed openstack/networking-ovn: Handle OVN NB and SB databases with separate ovsdb-server procs
*** jckasper has joined #openstack-neutron-ovn20:31
*** gsagie_ has quit IRC20:37
openstackgerritRyan Moats proposed openstack/networking-ovn: Handle OVN NB and SB databases with separate ovsdb-server procs
*** russellb_ has joined #openstack-neutron-ovn20:40
*** flaviof_ has joined #openstack-neutron-ovn20:40
*** russellb has quit IRC20:47
*** flaviof has quit IRC20:47
*** Sam-I-Am has quit IRC20:47
*** elurkki has quit IRC20:47
*** flaviof_ is now known as flaviof20:47
*** Sam-I-Am has joined #openstack-neutron-ovn20:49
*** elurkki has joined #openstack-neutron-ovn20:49
*** Sam-I-Am has quit IRC20:50
*** elurkki has quit IRC20:50
*** elurkki has joined #openstack-neutron-ovn21:03
*** jckasper has quit IRC21:07
*** armax has quit IRC21:12
openstackgerritMerged openstack/networking-ovn: Vagrant: Workaround shebang insanities
openstackgerritRyan Moats proposed openstack/networking-ovn: Handle OVN NB and SB databases with separate ovsdb-server procs
openstackgerritRyan Moats proposed openstack/networking-ovn: Split OVN SB from OpenVswitch databases
*** Sam-I-Am has joined #openstack-neutron-ovn21:35
*** roeyc has joined #openstack-neutron-ovn21:50
*** armax has joined #openstack-neutron-ovn21:53
openstackgerritMatthew Kassawara proposed openstack/networking-ovn: Devstack: Add native MTU option
Sam-I-Amrussellb_: moo?21:57
*** russellb_ is now known as russellb21:57
*** russellb has joined #openstack-neutron-ovn21:57
Sam-I-Amor, this russellb21:58
Sam-I-Amre: 27539221:58
russellb(been busy with topic:stadium-implosion)21:58
Sam-I-Ami noticed a lot of patches go by21:58
Sam-I-Amrussellb: so here's my thinking on whether or not to default dhcp and metadata agents on compute nodes....21:59
*** shettyg has quit IRC21:59
Sam-I-Ami think the gist is that people should test ovn with more than one dhcp and metadata agent, but its easier now (and more practical) to deploy >1 compute node than controller node, especially considering the db replication issues with ovn22:00
Sam-I-Amin a prod environment, people may choose to run dhcp/meta agents on multiple controllers, or maybe a handful of compute nodes, or maybe all compute nodes22:00
Sam-I-Amso i'm not really sure yet how to default this aside from hoping people read the docs and determine what to do22:01
Sam-I-Amfor example, with vagrant, we deploy the agents on compute nodes22:01
russellbok, so maybe a comment above the commented out enable_service lines that says something like ....22:01
regXboitwo more patches solder, just two more patches22:01
regXboier *soldier*22:02
*** jckasper has joined #openstack-neutron-ovn22:02
russellb"Feel free to enable these services on a compute node to test dhcp/metadata on more than one node. In a production env, you would not typically run these on every compute node, as you'd probably run them on multiple controllers or network nodes instead.  There's no harm in doing so for testing purposes.22:02
russellbor something22:02
russellband then i'll be happy22:02
russellbwhen it comes to configurable things, i have a philosophy22:02
russellb1) "make it work" options should die and not exist22:03
russellb2) have sane defaults22:03
russellb3) In every case something is optional, provide a clear explanation of why someone might want to change it22:03
Sam-I-Am"make it work" options? like make_neutron_work=true ?22:03
russellbi could come up with some real examples22:03
russellbthere's some stupid options out there that make no sense to ever disable22:04
russellbjust because someone thought that because it was "new" behavior, it should be optional22:04
russellbeven though the new behavior is the only sane thing22:04
russellbanyway :)22:04
Sam-I-Amwait, isnt that how neutron works?22:04
russellbit's how a lot of software works22:04
Sam-I-Amwrite feature, do not enable feature in gate, claim it works?22:04
Sam-I-Amrussellb: i'll rewrite the note22:05
Sam-I-Amand fix this merge conflict22:05
Sam-I-Amthis patch was the guinea pig for the tox stuff22:06
openstackgerritRyan Moats proposed openstack/networking-ovn: Refactor separate ovsdb-server launches
regXboiok, that completes that patch series22:09
*** roeyc has quit IRC22:11
*** roeyc has joined #openstack-neutron-ovn22:12
openstackgerritMatthew Kassawara proposed openstack/networking-ovn: Devstack: Improve DHCP and metadata agents
Sam-I-Amrussellb: maybe i got it this time22:21
openstackgerritlitong01 proposed openstack/networking-ovn: Vagrant: make provider network creation configurable
*** jckasper has quit IRC22:25
*** salv-orl_ has joined #openstack-neutron-ovn22:39
*** mamulsow has quit IRC22:42
*** salv-orlando has quit IRC22:42
arosen1russellb: I think i'm getting closer to tracking down what's causing this Unable to complete operation on network f4a75361-c158-44ae-9244-505c6e8b73e8. There are one or more ports still in use on the network.', u'detail': u''}22:49
arosen1Check this out:22:49
arosen1I added this log statement here in neutron and used Depends-On:
arosen1i'm wondering if it's a problem with the delete statement above here:
arosen1that we're getting dirty reads after the delete.22:52
*** rtheis has quit IRC22:54
Sam-I-Amarosen1: dhcp port sticking around?22:55
*** armax has quit IRC22:57
*** gangil has quit IRC23:00
*** chandrav has quit IRC23:02
*** gangil has joined #openstack-neutron-ovn23:03
*** gangil has joined #openstack-neutron-ovn23:03
*** allan_h has quit IRC23:05
*** allan_h has joined #openstack-neutron-ovn23:05
*** arosen1 has quit IRC23:07
*** Alik has joined #openstack-neutron-ovn23:18
*** regXboi has quit IRC23:19
Alikrussellb: Thanks Russell for your comments in respect to the topology where traversing through a few routers is required23:19
Alikrussellb: I'll appreciate if you confirm that multiple routers traversing topology is theoretically expected to work with the current network-ovn code (unless there are some bugs - I will dig into the OF rules more thoroughly) and elaborate how is this topology is being pushed from SB DB into the node, in terms of the general pipeline flow, it will help me to debug my case, if I know how the pipeline flow rules are expec23:20
*** igordcard_ has quit IRC23:26
*** aginwala has joined #openstack-neutron-ovn23:27
*** allan_h has quit IRC23:27
*** uday has joined #openstack-neutron-ovn23:46
*** azbiswas has quit IRC23:49
*** aginwala has quit IRC23:51
*** aginwala has joined #openstack-neutron-ovn23:56

Generated by 2.14.0 by Marius Gedminas - find it at!