Monday, 2017-07-17

*** kiennt has joined #openstack-kuryr00:35
*** caowei has joined #openstack-kuryr01:31
*** limao has joined #openstack-kuryr01:50
*** longfei_zhang has joined #openstack-kuryr01:57
*** kiennt has quit IRC02:07
openstackgerritNgo Quoc Cuong proposed openstack/fuxi master: Fix N529 Method's default argument should not be mutable  https://review.openstack.org/48202302:21
*** hyunsun has joined #openstack-kuryr02:28
*** yamamoto has joined #openstack-kuryr02:47
*** yamamoto has quit IRC03:09
*** s1061123 has joined #openstack-kuryr03:13
*** kiennt has joined #openstack-kuryr03:18
*** hongbin has joined #openstack-kuryr03:39
*** gouthamr has quit IRC03:42
*** kiennt has quit IRC03:52
*** kiennt has joined #openstack-kuryr03:59
*** yamamoto has joined #openstack-kuryr04:09
*** hongbin has quit IRC04:13
*** yamamoto has quit IRC04:14
*** limao has quit IRC04:16
*** tonanhngo has quit IRC05:10
*** limao has joined #openstack-kuryr05:13
*** longfei_zhang has quit IRC05:15
*** tonanhngo has joined #openstack-kuryr05:16
*** tonanhngo has quit IRC05:30
*** kiennt has quit IRC05:45
*** limao has quit IRC06:02
*** kiennt has joined #openstack-kuryr06:03
*** kiennt_ has joined #openstack-kuryr06:17
*** kiennt has quit IRC06:17
*** caowei has quit IRC06:45
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Skips kubernetes service without selector  https://review.openstack.org/48389106:48
*** janki has joined #openstack-kuryr06:49
ltomasboirenab, you are really fast! :D06:52
irenabltomasbo: haha, I just ran gerrit query for open patches and your came up first. Now I see you just uploaded it06:53
ltomasbo:D06:54
ltomasboyep, like 20 seconds before your review!06:54
ltomasboI'll push a new patch in a minute!06:54
*** kzaitsev_ws has joined #openstack-kuryr06:55
irenabpk06:56
irenabok06:56
openstackgerritMerged openstack/fuxi master: Fix N529 Method's default argument should not be mutable  https://review.openstack.org/48202306:58
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Skips kubernetes service without selector  https://review.openstack.org/48389107:01
*** tonanhngo has joined #openstack-kuryr07:02
*** hyunsun has quit IRC07:03
*** tonanhngo has quit IRC07:03
openstackgerritMerged openstack/fuxi master: Use https instead of http in docs links  https://review.openstack.org/45440807:08
openstackgerritMerged openstack/fuxi master: Move the imports to the top of the file  https://review.openstack.org/47563407:08
openstackgerritMerged openstack/fuxi master: Enable some off-by-default checks  https://review.openstack.org/47636207:08
*** caowei has joined #openstack-kuryr07:09
openstackgerritJaivish Kothari(janonymous) proposed openstack/kuryr-kubernetes master: [DRAFT] CNI Daemon template  https://review.openstack.org/48002807:10
janonymousirenab: ^^ , i am still facing some issues with vif allocation regarding resource busy.. :(07:11
irenabjanonymous: did you try deployment without your change? I mean are you sure the issue is related?07:12
janonymousirenab: ahh, no i got errors without patch also..some race condition i am not sure about07:13
irenabBM or nested?07:14
janonymousBM07:14
irenabI will try to deploy as well07:14
janonymousthanks!07:15
openstackgerritMerged openstack/kuryr-tempest-plugin master: Updated from global requirements  https://review.openstack.org/48341107:19
*** aojea has joined #openstack-kuryr07:19
*** egonzalez has joined #openstack-kuryr07:40
openstackgerritNgo Quoc Cuong proposed openstack/fuxi master: Reuse neutron-lib for pep8 hacking checks  https://review.openstack.org/48427607:52
kzaitsev_wsjanonymous: it keeps saying that reourse is busy, untill timeouts?07:57
kzaitsev_wsjanonymous: check if the relevant agent is running07:58
kzaitsev_wsneutron-openvswitch-agent for example.07:59
kzaitsev_wswe really need to start a FAQ for this purporse =)08:01
janonymouskzaitsev_ws:  yes, it was running when i checked08:07
janonymouskzaitsev_ws:  yeah :)08:08
janonymouskzaitsev_ws: Did you faced similar issue ?08:08
kzaitsev_wsjanonymous: from time to time.08:08
kzaitsev_wsfor me it usually means that there is a problem in agent-neutron communication.08:09
kzaitsev_wsyou can check agent logs for timeouts. and also check if rabbit is feeling ok. and if neutron thinks agent is up (neutron agent-list)08:10
janonymouskzaitsev_ws:  All these things were looking good, and did you also got commit problem with vif08:11
kzaitsev_wscommit problem with vif?08:12
janonymousyes, ipdb when it tries to call commit()08:12
kzaitsev_wsah. no, never seen that08:13
kzaitsev_wswhat does it say?08:13
janonymousi will prepare a new env from scratch again... unable to commit..08:14
janonymousi don't have logs i guess now08:14
janonymouswill paste when reoccurs..08:15
janonymouskzaitsev_ws: thanks for the pointers08:20
kzaitsev_wsjanonymous: you can also check if ovs is running/configured correctly and that it has the port added br-int (or whatever is your integration bridge name)08:25
kzaitsev_wsyou should probably start with it %) if ovs doesn't work — there's no point in checking agent/neutron/rabbit %)08:26
kzaitsev_wss/added/added to/08:26
irenabjanonymous: I deployed BM and checked k8s guestbook all in one app, I see all pods are running08:33
janonymousirenab: kzaitsev_ws : i was using running simple vagrant ...`vagrant up` to run VM with devstack08:34
janonymousirenab: Can i know the step you used08:35
irenabjanonymous: same here. But once I entered the VM, I saw devstack failed, so did unstack and stack again08:35
irenabthen it completed08:35
janonymous:D08:36
janonymousirenab:  can you try `kubectl run hello-world --replicas=2 --labels="run=load-balancer-example" --image=gcr.io/google-samples/node-hello:1.0  --port=8080`08:37
janonymousirenab:  and see if you can find errors in logs08:37
janonymousirenab:  remove  --labels option08:37
irenabalready posted with the label…08:38
irenabwhat log to check? kubelet?08:38
irenabboth replicas seemsto be running08:39
irenabjanonymous: ping from one instance to another is working08:41
janonymousirenab:  yes, kuryr-kubernetes and kubelet logs08:41
irenabjanonymous: shall I look for something specific?08:42
*** garyloug has joined #openstack-kuryr08:42
*** garyloug has quit IRC08:42
irenabkuryr-kubernetes seems fine08:43
janonymousirenab:  during pod ADD call , something after: http://paste.openstack.org/show/615582/08:43
irenabkubelet shows ‘ failed to handle multiple devices for container. Skipping Filesystem stats’, but networking seems to be working and pods are running08:43
janonymousirenab: maybe something i am missing / doing wrong..i will recheck with fresh devstack setup and update shortly08:44
irenabjanonymous: Do you still run devstack with screen option?08:47
irenabI run the default (systemd). Maybe it can be related08:47
janonymousirenab: yup, with USE_SCREEN=True08:48
janonymousirenab: ahh, maybe08:48
irenabjanonymous: try systemd, I am not sure how reliable the screen version08:48
janonymousirenab: oh, ohkay08:48
irenabdmellado: did you try screen devstack version lately?08:49
*** garyloug has joined #openstack-kuryr08:49
dmelladoirenab: not really, IIRC there were some issues with that08:49
dmelladowhat's broken this time? :D08:49
irenabjanonymous: ^^08:49
janonymousirenab:  ohh..08:49
dmelladojanonymous: if you want to have a look at that low-hanging-fruit it'd be great!:P08:50
janonymousdmellado:  Haha :P08:50
janonymousirenab: kzaitsev_ws : dmellado : i will try with systemd now, last time i tried systemd i was stuck with log processing using journalctl... and restart issue of kuryr-kubernetes which i had to do manually08:52
*** pmannidi has joined #openstack-kuryr08:58
irenabjanonymous: I think dmellado is the master of the systemd tricks, maybe he can advise on logs and restart09:03
dmelladojanonymous: do you mean in the current desvstack? it just work for me09:04
dmelladoI was putting up quite a few patches for that, mainly stupid 'sudos' around09:04
dmelladoso pls do give it a try and ping me if there's any issues09:04
janonymousirenab:  :)09:05
janonymousdmellado:  It started pretty good for me, but running a pod had some few errors. did you tried vagrant lately?09:06
dmelladojanonymous: which vagrant do you mean? I usually just use the local.conf within rdo cloud ;)09:08
dmelladodsvm, I mean09:08
janonymousdmellado:  haha, no it is local.conf.sample one i guess, but just clone kuryr-kubernetes and `cd contrib/vagrant` and `vagrant up` and `kubectl run hello-world --replicas=2  --image=gcr.io/google-samples/node-hello:1.0  --port=8080`09:11
dmelladojanonymous: let me have a quick look, it should just be the same09:11
janonymousdmellado: just add USE_SCREEN=True in local.conf.sample after cloning09:11
dmelladooh, I'd bet SCREEN won't work09:12
dmelladoIIRC there were some issues after we sorted out the systemd migration09:12
janonymousirenab:  you were right, it worked well without screen09:12
janonymousdmellado:  yes09:12
dmelladojanonymous: so feel free to submit a 'bug' and put it on the backlog09:12
*** yamamoto has joined #openstack-kuryr09:13
*** yamamoto has quit IRC09:13
*** yamamoto has joined #openstack-kuryr09:14
janonymousdmellado:  i will try to find the cause :)09:14
dmelladojanonymous: awesome! I'm pretty sure it should be quite the 'stupid' thing... but just didn't have the time to go into that, so thanks!09:15
*** yamamoto has quit IRC09:16
janonymousdmellado:  yup,np..have to resume my work...which i was thinking my patch issue :D09:17
*** yamamoto has joined #openstack-kuryr09:17
janonymousthanks guys!09:18
*** yamamoto has quit IRC09:21
*** yamamoto has joined #openstack-kuryr09:21
*** yamamoto has quit IRC09:26
*** yamamoto has joined #openstack-kuryr09:30
*** yamamoto has quit IRC09:38
kzaitsev_wsirenab: may I ask you to merge this one? https://review.openstack.org/#/c/481204/09:46
*** caowei has quit IRC10:11
*** yamamoto has joined #openstack-kuryr10:37
*** kiennt_ has quit IRC10:37
*** tonanhngo has joined #openstack-kuryr11:04
*** tonanhngo has quit IRC11:05
openstackgerritKirill Zaitsev proposed openstack/kuryr-kubernetes master: Allow passing multiple VIFs to CNI  https://review.openstack.org/47101211:09
*** atoth has quit IRC11:10
*** rwallner has joined #openstack-kuryr11:13
*** dulek has joined #openstack-kuryr11:18
dulekHi! I'm trying to run kuryr-k8s on OpenStack-Helm. I'm getting this traceback from kuryr-cni: http://paste.openstack.org/show/615595/11:21
dulekAny ideas about the cause?11:21
dulekOh, and I'm using https://review.openstack.org/#/c/46667511:21
*** aojea has quit IRC11:57
*** atoth has joined #openstack-kuryr12:03
kzaitsev_wsdulek: you should probably ping vikasc ;)12:11
openstackgerritKirill Zaitsev proposed openstack/kuryr-kubernetes master: Allow passing multiple VIFs to CNI  https://review.openstack.org/47101212:11
*** pmannidi has quit IRC12:13
dulekHi vikasc! :)12:14
vikascdulek, never faced this issue. Thinking what could be the reson :)12:14
vikascs/reason12:14
vikascdulek, hi12:14
dulekvikasc: I'm running my k8s on Ubuntu 16.04. You've probably tested it only on RH/CentOS?12:15
vikascdulek, yes, thats right. but since cni is a standalone executable finally, it should have worked12:16
vikascdulek, i will have to try on ubuntu, to get anyclue12:17
dulekvikasc: https://github.com/svinota/pyroute2/blob/master/pyroute2/ipdb/interfaces.py#L4012:18
dulekvikasc: Maybe bond data is somehow unaccessible on 16.04.12:18
vikascdulek, pyroute2 is installed in cni builder container and then using pyinstaller, cni executable is created and copied on host namespace12:19
dulekvikasc: But it would be strange for pyroute2 to be not portable.12:19
vikascdulek, cni is build inside a centos container only12:20
vikascs/built12:20
vikasceven for ubuntu host12:20
vikascdulek, https://review.openstack.org/#/c/466675/14/cni_builder.Dockerfile@112:21
*** jermz has left #openstack-kuryr12:21
dulekvikasc: Okay… So I still have no clue what's wrong…12:23
vikascdulek, i will try it and get back12:23
dulekvikasc: Thanks!12:24
vikascdulek, no problem, i will getback12:25
vikascdulek, thanks for trying this out12:26
irenabkzaitsev_ws: done12:49
*** aojea has joined #openstack-kuryr12:52
openstackgerritMerged openstack/kuryr master: Updated from global requirements  https://review.openstack.org/48120412:53
*** rwallner has quit IRC12:56
*** aojea has quit IRC12:56
openstackgerritDaniel Mellado proposed openstack/kuryr-kubernetes master: Add a flag for enabling trunk in neutron  https://review.openstack.org/48436313:00
*** gouthamr has joined #openstack-kuryr13:01
*** rwallner_ has joined #openstack-kuryr13:01
openstackgerritDaniel Mellado proposed openstack/kuryr-kubernetes master: Add a flag for enabling trunk in neutron  https://review.openstack.org/48436313:07
kzaitsev_wsirenab: ok, now we need apuimedo to make a release of kuryr-lib =)13:08
kzaitsev_wsor I can request it and add it apuimedo to the reviewers13:08
*** aojea has joined #openstack-kuryr13:19
*** aojea has quit IRC13:26
*** kzaitsev_ has joined #openstack-kuryr13:30
dmelladoapuimedo: ping13:48
*** aojea has joined #openstack-kuryr13:48
dmelladokzaitsev_ws: +1 to the 2nd option xD13:48
ltomasbodmellado, apuimedo is on PTO13:48
dmelladooh, I see, I forgot about it13:48
dmelladokzaitsev_ws: btw, I was having a look at your gist at https://github.com/openstack/kuryr-kubernetes/blob/master/doc/source/installation/nested-vlan.rst13:49
*** yamamoto has quit IRC13:49
dmelladoand I wanted to discuss how to automate this for the upstream gate13:49
dmelladoI was speaking about it with irenab this morning13:49
dmelladoand I put a patch before for this13:50
dmelladoltomasbo: you were telling me that apuimedo already started something related to this effort, didn't you?13:52
dmelladonext time I'll force him to put up a proper blueprint...13:52
ltomasbodmellado, not exactly that13:55
ltomasbohttps://review.openstack.org/#/c/472763/13:55
dmelladooh, true, I forgot about that patch13:56
* dmellado has just too many things on that13:56
dmelladoirenab: hope you're feeling better, were you using also that on your efforts for df?13:56
janonymousMeeting today?13:58
dmelladoLet's hop in there, sure ;)14:00
*** tonanhngo has joined #openstack-kuryr14:01
*** tonanhngo has quit IRC14:02
*** kzaitsev_ has quit IRC14:08
*** yamamoto has joined #openstack-kuryr14:09
*** yamamoto has quit IRC14:09
*** hongbin has joined #openstack-kuryr14:11
openstackgerritDanil Golov proposed openstack/kuryr-kubernetes master: Allow setting specific ports for SRIOV handler  https://review.openstack.org/47849414:21
openstackgerritDanil Golov proposed openstack/kuryr-kubernetes master: Allow setting specific ports for SRIOV handler  https://review.openstack.org/47849414:30
*** zengchen1 has joined #openstack-kuryr14:33
zengchen1hi, guys. is the weekly meeting canceled? i don't see any talks. thanks.14:38
*** yamamoto has joined #openstack-kuryr14:40
*** aojea has quit IRC14:40
*** dell_ has joined #openstack-kuryr14:41
*** zengchen1 has quit IRC14:44
*** yamamoto has quit IRC14:46
*** dell_ is now known as zengchen114:48
*** janki has quit IRC15:02
dmelladozengchen1: o/ we were done quickly today15:04
dmelladoit was on openstack-meeting-4 but we're already done15:04
dmelladodid you want to bring any topic?15:04
openstackgerritDanil Golov proposed openstack/kuryr-kubernetes master: Allow setting specific ports for SRIOV handler  https://review.openstack.org/47849415:19
*** kzaitsev_ has joined #openstack-kuryr15:25
*** rwallner_ has quit IRC15:25
*** kzaitsev_ has quit IRC15:27
*** rwallner has joined #openstack-kuryr15:30
*** aojea has joined #openstack-kuryr15:31
*** kzaitsev_ has joined #openstack-kuryr15:31
*** kzaitsev_ has quit IRC15:34
*** aojea has quit IRC15:36
*** zengchen1 has quit IRC15:36
kzaitsev_wsdmellado: sorry I got dragged to a meeting and couldn't say no =(15:48
kzaitsev_wsirenab: garyloug: would be happe to review any doc/code on the multiple networks topic15:51
kzaitsev_wsI've put up a small patch that implements a naïve approach to that15:51
*** atoth has quit IRC15:53
kzaitsev_wsthat is I need them sooner rather than later. Maybe (if developing the right, TPR/CRD approach takes a lot of time) we can implement phased approach here. mark multi-network as experimental or smth and upgrade to a more robust implementation after15:55
kzaitsev_wsor if the code is up soon — we can just use that =) I'd be happy )15:55
kzaitsev_wsgaryloug: btw. I believe we talked about using TPR now and upgrading to CRD later15:56
*** egonzalez has quit IRC16:02
kzaitsev_wsdmellado: btw doc ^^ is not really mine. I just split aq single huge README into multiple small docs and put them under toc in the documentation. I wrote the manual installation piece though16:05
kzaitsev_wsI also have to update the thing to comply with the big-unified-doc-structure %)16:06
dmelladokzaitsev_ws: ack, did you get to read today's notes?16:11
*** atoth has joined #openstack-kuryr16:21
kzaitsev_wsyeah, looked through. anything I should pay attention to?16:22
*** kural has joined #openstack-kuryr16:23
kuralkzaitsev_ws hi16:23
kzaitsev_wsapuimedo et al.: https://review.openstack.org/#/c/484416/ — requesting a 0.5.0 release and https://review.openstack.org/484420 requesting a requirements bump.16:23
kzaitsev_wskural: hi. anything I can help you with? =)16:24
kuralRegarding the CRD16:24
kuralTPR will be removed from 1.8 K8s16:24
kuralSo it is better to do POC work in CDR..16:24
kuralSelflink of the Kube-API for both TPR and CDR should be the same, So all the users less than 1.7 can use TPR, others(>=1.8) can use CDR16:26
*** aojea has joined #openstack-kuryr16:26
kuralwe should be backward compatible in our approach16:27
kzaitsev_wskural: if they're that compatible — great. =)16:29
kzaitsev_wsit's just that install base for k8s 1.7 is probably non-existant )16:29
*** aojea has quit IRC16:30
kzaitsev_wsand turning kuryr-k8s incompatible with <=1.7 doesn't sound like a good idea16:31
kzaitsev_wsbut if you're saying that implementation would be compatible — that's great16:31
*** garyloug has quit IRC16:36
*** aojea has joined #openstack-kuryr16:45
*** aojea has quit IRC16:49
*** kzaitsev_ws has quit IRC16:50
*** aojea has joined #openstack-kuryr17:03
*** aojea has quit IRC17:07
*** aojea has joined #openstack-kuryr17:12
*** aojea has quit IRC17:16
*** aojea has joined #openstack-kuryr17:21
*** aojea has quit IRC17:25
*** tonanhngo has joined #openstack-kuryr17:28
*** aojea has joined #openstack-kuryr17:31
*** tonanhngo has quit IRC17:32
*** aojea has quit IRC17:36
*** tonanhngo has joined #openstack-kuryr17:38
*** aojea has joined #openstack-kuryr17:40
*** aojea has quit IRC17:45
*** aojea has joined #openstack-kuryr17:50
*** aojea has quit IRC17:51
*** aojea has joined #openstack-kuryr17:51
*** aojea has quit IRC17:56
*** aojea has joined #openstack-kuryr18:07
*** aojea has quit IRC18:11
*** aojea has joined #openstack-kuryr18:16
*** aojea has quit IRC18:21
*** aojea has joined #openstack-kuryr18:35
*** aojea has quit IRC18:37
*** rwallner has quit IRC18:51
*** kural has quit IRC19:26
*** rwallner has joined #openstack-kuryr20:00
*** rwallner has quit IRC20:00
*** rwallner has joined #openstack-kuryr20:01
*** rwallner has quit IRC20:08
*** sridhar_ram has joined #openstack-kuryr20:33
dimsapuimedo : don't see you on #openstack-release... can you please bless release request https://review.openstack.org/#/c/484416/ from kzaitsev_pi ?20:34
*** kural has joined #openstack-kuryr20:54
*** kural has quit IRC21:03
*** kural_ has joined #openstack-kuryr21:03
openstackgerritOpenStack Proposal Bot proposed openstack/fuxi master: Updated from global requirements  https://review.openstack.org/48352222:00
*** gouthamr has quit IRC22:12
*** gouthamr has joined #openstack-kuryr22:15
*** apuimedo has quit IRC22:16
*** apuimedo has joined #openstack-kuryr22:17
*** tonanhngo has quit IRC22:20
*** tonanhngo has joined #openstack-kuryr22:22
*** tonanhngo_ has joined #openstack-kuryr22:26
*** tonanhngo has quit IRC22:26
*** tonanhngo_ has quit IRC22:30
*** kural_ has quit IRC22:36
*** hongbin has quit IRC23:28
*** gouthamr has quit IRC23:32
*** gouthamr has joined #openstack-kuryr23:51
*** gouthamr has quit IRC23:53
*** gouthamr has joined #openstack-kuryr23:54

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!