Wednesday, 2016-11-30

« Tuesday, 2016-11-29 Index Thursday, 2016-12-01 »
*** devlaps has joined #openstack-kuryr00:02
*** hongbin has quit IRC00:20
*** diogogmt has joined #openstack-kuryr00:24
*** limao has joined #openstack-kuryr00:37
apuimedodougbtv: we're starting to work on openshift on openstack00:40
apuimedowith kuryr00:40
apuimedodougbtv: did you mean side by side in your question? With both openstack and openshift on baremetal00:40
apuimedoor wih openshift running on nova instances?00:40
*** tonanhngo has quit IRC00:50
*** tonanhngo has joined #openstack-kuryr00:51
*** tonanhngo has quit IRC00:55
*** ApheleiaS has quit IRC01:12
*** apuimedo is now known as apuimedo|away01:42
openstackgerritAntoni Segura Puimedon proposed openstack/kuryr-libnetwork: requirements: bump to kuryr-lib 0.2.0  https://review.openstack.org/40449701:44
*** shashank_hegde has quit IRC02:06
*** devlaps has quit IRC02:07
*** hongbin has joined #openstack-kuryr02:21
dougbtvapuimedo|away, thanks btw! I'd be interested in both. Right now I have openstack on baremetal and openshift in nova instances as a reference/in-a-lab, so I guess I'm interested in that first. But, side-by-side both on baremetal is also interesting to me. I'd be interested in lending a hand / contributing to that, too. Hopefully catch up another time!!02:29
*** diogogmt has quit IRC02:31
*** hongbin has quit IRC02:32
*** yuanying has quit IRC02:46
*** yuanying has joined #openstack-kuryr02:55
janonymousalraddarla_: hi03:08
*** pmannidi has quit IRC03:09
*** shashank_hegde has joined #openstack-kuryr03:28
*** vikasc has joined #openstack-kuryr03:35
vikascapuimedo|away, was not well yesterday, so was away03:37
*** yuanying has quit IRC03:49
*** yuanying has joined #openstack-kuryr04:18
*** yamamoto_ has quit IRC04:24
*** hongbin has joined #openstack-kuryr04:30
*** hongbin has quit IRC04:30
*** pmannidi has joined #openstack-kuryr04:39
*** shashank_hegde has quit IRC04:48
*** jgriffith is now known as jgriffith_away04:54
*** shashank_hegde has joined #openstack-kuryr05:08
*** yamamoto_ has joined #openstack-kuryr05:08
*** shashank_hegde has quit IRC05:15
*** saneax-_-|AFK is now known as saneax05:19
*** shashank_hegde has joined #openstack-kuryr05:50
*** shashank_hegde has quit IRC05:54
*** yedongcan has joined #openstack-kuryr06:31
*** shashank_hegde has joined #openstack-kuryr06:32
openstackgerritDongcan Ye proposed openstack/kuryr-libnetwork: Fix bug in ipam_release_address  https://review.openstack.org/40187406:44
*** tonanhngo has joined #openstack-kuryr06:46
*** oanson has joined #openstack-kuryr07:01
openstackgerritDongcan Ye proposed openstack/kuryr-libnetwork: Cleanup remove_container method in Rally job  https://review.openstack.org/40459107:13
*** irenab has joined #openstack-kuryr07:20
*** shashank_hegde has quit IRC07:24
*** shashank_hegde has joined #openstack-kuryr07:24
*** tonanhngo has quit IRC07:38
*** shashank_hegde has quit IRC07:41
*** pmannidi has quit IRC07:43
*** tonanhngo has joined #openstack-kuryr07:44
*** tonanhngo has quit IRC07:46
*** shashank_hegde has joined #openstack-kuryr07:47
*** yedongcan has quit IRC07:52
apuimedo|awayvikasc: me too07:55
apuimedo|awayI had a pretty nasty tonsilitis07:55
*** apuimedo|away is now known as apuimedo07:55
*** tonanhngo has joined #openstack-kuryr07:57
*** yedongcan has joined #openstack-kuryr07:58
*** tonanhngo has quit IRC07:58
openstackgerritAntoni Segura Puimedon proposed openstack/kuryr-kubernetes: requirements: bump to kuryr-lib 0.2.0  https://review.openstack.org/40461607:58
apuimedovikasc: irenab: kuryr kubernetes still does not use the reqs bot. Let's use ^^08:03
*** limao has quit IRC08:03
*** shashank_hegde has quit IRC08:08
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr-libnetwork: Nested-Containers: trunk subports management  https://review.openstack.org/40246208:11
*** dimak has joined #openstack-kuryr08:16
ltomasboirenab: I addressed your comments, please take a look when you have time08:17
irenabltomasbo: sure, asap08:18
ltomasbothanks!08:18
*** limao has joined #openstack-kuryr08:19
yedongcanlimao: ping08:26
limaoyedongcan: hello08:29
yedongcanlimao: Sometimes, rally job will failed, error is "Unable to remove filesystem"08:30
yedongcanlimao: I think this is a bug of docker, and remove container is unrelated with network, do we need to remove container in our rally job?08:31
limaoyedongcan: yes, I aslo see it, and find some bug in docker08:31
yedongcanlimao: I update a patch for this, currently we can cleanup the method in rally job, how do you think about?08:34
limaoyedongcan: I think we should clean it up since we create it :-)  You mean you want to only stop it right?08:34
limaoyedongcan: can you paste the link?08:34
limaoyedongcan: ohh, I see it  https://review.openstack.org/40459108:35
yedongcanlimao: yes, we only stop it.08:35
limaoyedongcan: I think it is OK, maybe we should just comment it out(not remove), and leave some comments and bug link there08:36
limao(I mean in the patch, once it is fixed in docker, it should be ok we remove the container, what do you think?)08:36
yedongcanlimao: agree with you08:38
limaoyedongcan : Thanks for this!  and hopefully I will update the patch https://review.openstack.org/#/c/396048/ this week, so that it did not block you, thanks08:39
limao(should be https://review.openstack.org/#/c/387835/)08:39
*** tonanhngo has joined #openstack-kuryr08:41
yedongcanlimao::-D thanks08:41
irenabltomasbo: ping08:41
ltomasboirenab: pong08:41
*** shashank_hegde has joined #openstack-kuryr08:43
*** tonanhngo has quit IRC08:43
irenabltomasbo: I checked the new patch. I am not sure I understand what you are trying to do there with drivers. My suggestion was that you split the Bare Metal and Nested handling into separate driver that controller will invoke for libnetwork operations. So this way you can keep the main Controller generic and hold specific handling inside each driver08:43
ltomasbowith baremetal you mean veth?08:44
irenabltomasbo: probably yes08:44
irenabthe current support08:45
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr-libnetwork: Nested-Containers: trunk subports management  https://review.openstack.org/40246208:45
irenabIdeally, Controller should stay unchanged when some changes applied on nested/veth special handling or when some other type is added08:46
*** lezbar has joined #openstack-kuryr08:46
ltomasboshould that be the same patch? or a different one where the functionality is split?08:46
ltomasbolet's take the example of ipam_request_address() function08:47
irenabltomasbo: if separate patch, it nested support should depend on ‘split controller and introduce drivers’ patch08:47
irenabltomasbo: you just need to identify where are the points that support is different and delegate handling of this to drivers. All the common parts should be in Controller08:48
*** lezbar has quit IRC08:48
ltomasboaren't the common parts already in the controller08:49
ltomasboand I just created some functions based on the driver being used08:49
*** lezbar has joined #openstack-kuryr08:49
*** diogogmt has joined #openstack-kuryr08:49
ltomasbodo you want those in a different .py?08:49
irenabltomasbo: my problem now that non the common are added into the controller08:49
ltomasbosuch as?08:50
irenabltomasbo: different .py is the way to go, but its more than this. When ‘veth’ driver is used, it should not even load Seg.Driver or check for the ‘vlan aware vm’ support08:50
ltomasbobut, with the last patch, when the driver is veth, Seg.Driver is not loaded08:51
ltomasboonly when the driver is vlan08:51
ltomasboand the same for the check_for_vlan_ids08:52
ltomasbothe load_driver() methods calls load_driver_veth() when driver is veth (which does nothing)08:52
ltomasboand calls the load_driver_vlan() when driver is vlan08:52
ltomasbowhich loads the seg_driver plus do the vlan_ids checking08:52
*** shashank_hegde has quit IRC08:53
ltomasboirenab: look at lines 90 - 107 of the new patch set08:53
*** lezbar has quit IRC08:54
irenabltomasbo: I see it now … import in the middle … Please split each driver to different .py, it will help with the Controller readability08:54
irenabltomasbo: checking08:54
*** tonanhngo has joined #openstack-kuryr08:54
*** tonanhngo has quit IRC08:55
irenabltomasbo: just realised that Controller is more than 1600 lines! Please split the driver specifics08:55
ltomasboyes, that I agree, but not sure if that should be in the same patch set or not08:56
ltomasboas the purpose of it is different08:56
irenabltomasbo: I do not mind if you do it as a step before ading nested case08:56
ltomasboone is include support to vlan, and the other is to make controller.py driver plugable (to put it that way)08:56
irenabltomasbo: the opposite order08:56
irenabbut I think you may find it easier to handle it all together08:58
openstackgerritDongcan Ye proposed openstack/kuryr-libnetwork: Temporary cleanup remove_container method in Rally job  https://review.openstack.org/40459108:58
ltomasbook08:59
*** diga has joined #openstack-kuryr08:59
*** tonanhngo has joined #openstack-kuryr09:13
*** tonanhngo has quit IRC09:15
*** lezbar has joined #openstack-kuryr09:22
*** garyloug has joined #openstack-kuryr09:40
*** yedongcan has left #openstack-kuryr09:49
*** tonanhngo has joined #openstack-kuryr09:53
*** tonanhngo has quit IRC09:54
*** janki has joined #openstack-kuryr10:07
*** lmdaly has joined #openstack-kuryr10:11
*** tonanhngo has joined #openstack-kuryr10:11
*** tonanhngo has quit IRC10:12
*** fkautz has quit IRC10:19
*** fkautz has joined #openstack-kuryr10:21
*** garyloug_ has joined #openstack-kuryr10:23
*** janonymous_ has joined #openstack-kuryr10:23
*** limao_ has joined #openstack-kuryr10:25
*** janonymous has quit IRC10:25
*** garyloug has quit IRC10:25
*** limao has quit IRC10:25
*** janonymous_ is now known as janonymous10:25
*** limao_ has quit IRC10:27
*** diga has quit IRC10:28
*** tonanhngo has joined #openstack-kuryr10:30
*** tonanhngo has quit IRC10:32
*** pcaruana has joined #openstack-kuryr10:35
*** tonanhngo has joined #openstack-kuryr10:44
*** tonanhngo has quit IRC10:44
*** tonanhngo has joined #openstack-kuryr10:57
*** tonanhngo has quit IRC10:58
*** irenab has quit IRC10:58
*** janki has quit IRC11:05
*** tonanhngo has joined #openstack-kuryr11:10
*** tonanhngo has quit IRC11:10
*** tonanhngo has joined #openstack-kuryr11:23
*** tonanhngo has quit IRC11:24
*** gsagie has joined #openstack-kuryr11:33
*** irenab has joined #openstack-kuryr11:35
*** gsagie has quit IRC11:36
*** diogogmt has quit IRC12:15
mchiapperoa quick one12:15
*** diogogmt has joined #openstack-kuryr12:15
mchiapperowhen documenting the thrown exception, do you use the absolute path or the one imported in that file?12:16
mchiapperoI saw a mix12:16
*** limao has joined #openstack-kuryr12:24
*** lmdaly has quit IRC12:26
*** tonanhngo has joined #openstack-kuryr12:28
*** tonanhngo has quit IRC12:29
*** diogogmt has quit IRC12:33
*** garyloug_ has quit IRC12:45
*** yamamoto_ has quit IRC12:59
*** tonanhngo has joined #openstack-kuryr13:13
*** tonanhngo has quit IRC13:14
apuimedomchiappero: what does 'documenting the thrown exception' mean?13:39
*** oanson has quit IRC13:43
*** yamamoto has joined #openstack-kuryr13:45
mchiapperoI mean when documenting functions13:46
apuimedoyou mean inline doc?13:47
*** tonanhngo has joined #openstack-kuryr13:54
*** tonanhngo has quit IRC13:56
mchiapperoyes, sorry :)14:02
apuimedoimported one14:09
*** tonanhngo has joined #openstack-kuryr14:15
*** tonanhngo has quit IRC14:15
*** dougbtv has quit IRC14:22
*** lmdaly has joined #openstack-kuryr14:22
*** tonanhngo has joined #openstack-kuryr14:34
*** tonanhngo has quit IRC14:39
*** irenab has quit IRC14:40
*** garyloug_ has joined #openstack-kuryr14:41
*** tonanhngo has joined #openstack-kuryr14:41
*** tonanhngo has quit IRC14:42
*** devlaps has joined #openstack-kuryr14:53
*** saneax is now known as saneax-_-|AFK14:58
apuimedoivc_: http://paste.openstack.org/show/590967/14:59
apuimedothis is with kuryr-lib 0.2.014:59
ivc_apuimedo, thats ok :) it requires kuryr.lib 0.2.014:59
apuimedodo prdele14:59
apuimedoI had put in the requirements.txt in kuryr-kubernetes to use kuryr.lib 0.2.015:00
apuimedoand I have reclone=no15:00
apuimedoFFFFFFFFFUUUUU15:00
ivc_you also need to update upper-constraints :)15:00
apuimedolet's see what happened15:00
*** yamamoto has quit IRC15:00
apuimedoright, it is getting merged as we speak15:00
ivc_your patch to openstack/requirements has not been merged yet15:00
apuimedo:-)15:00
apuimedohttps://review.openstack.org/#/c/404619/15:01
*** tonanhngo has joined #openstack-kuryr15:01
ivc_cool :)15:01
*** tonanhngo has quit IRC15:01
ivc_apuimedo, i'm also trying to get devstack plugin fixed. it seems that 'prepare_kubernetes_files' does not work the same way in 1.4 as it did in 1.315:02
apuimedoivc_: yeah... irenab is having trouble with that15:02
apuimedoI guess I didn't stumble on that since I'm using the same volume15:03
apuimedoivc_: leave it up to me15:03
apuimedoI can get it fixed15:03
ivc_ok15:03
apuimedoivc_: I prefer you work on CNI driver :-)15:03
apuimedoif you don't mind15:03
ivc_yeah sure15:04
apuimedocool15:04
apuimedoivc_: I'm also checking about the ovs internal ports15:04
ivc_just want to run it in the same env others do, because otherwise some things slip away (i.e. my cni patch is missing few things as those were done as 'hacks' earlier on my env)15:05
*** yamamoto has joined #openstack-kuryr15:06
ivc_apuimedo, i've tried moving ovs internal ports into container's namespace and it did seem to work. tho maybe there are some subtle issues i've not found out yet15:07
apuimedoright15:08
apuimedothat's exactly why I'm checking :-D15:08
ivc_apuimedo, regarding https://review.openstack.org/#/c/404038/ i think we agreed that we need a daemon+exec but we also need some 'stub' before we get there15:09
apuimedostub is fine15:09
ivc_apuimedo, so i wonder if it is worth the effort to do unit-tests for https://review.openstack.org/#/c/404038/15:09
apuimedoand I'm still tempted to add file management15:09
apuimedoivc_: prolly only for the exec part15:09
ivc_no need for now. its a stub :)15:09
apuimedoso minimal unit tests15:10
ivc_but it is not split into daemon/exec yet15:10
ivc_and the daemon/exec impl will not use most of the code from this one15:11
apuimedothat's why I say to only unit test the basic interface15:11
apuimedowhich I expect to survive15:11
apuimedothe frontend15:11
ivc_what do you call 'the frontend'?15:11
apuimedothe exec part15:12
apuimedoso no unit tests for the watch part15:12
ivc_CNIRunner you mean?15:12
ivc_but its API will change for daemon/exec15:12
ivc_what is likely to survive the move to daemon+exec is the binding part15:13
*** oanson has joined #openstack-kuryr15:13
apuimedook15:13
apuimedoso let's keep it light for now15:14
apuimedoas usual, put it in the commit msg15:14
ivc_so, no tests for now?15:14
*** tonanhngo has joined #openstack-kuryr15:14
*** tonanhngo has quit IRC15:15
apuimedowhatever won't survive, I can live with having lack of tests briefly15:16
apuimedoas long as we can add functional/fullstack tests soon15:16
ivc_i just want to get my hands on daemon+exec ASAP, but for that i need to finish service/lbaas first15:16
apuimedoivc_: agreed on both accounts15:17
apuimedowe're aligned on the goals and order15:17
ivc_ok, cool15:17
apuimedoI finish checking the internal ports and move to the devstack config file generation15:19
*** limao has quit IRC15:27
openstackgerritDarla Ahlert proposed openstack/kuryr-libnetwork: Unittests with mock  https://review.openstack.org/39432415:29
apuimedoivc_: the internal ports seem to work15:35
apuimedoI put them on docker containers15:35
apuimedoboth for container to container and for container to VM15:38
apuimedomchiappero: ltomasbo: did you consider using ovs internal ports for the containers15:38
apuimedoand for vlan trunk just setting the vlan tag for the internal  port?15:39
apuimedoto avoid using macvlan/ipvlan15:39
ltomasbohi apuimedo15:41
apuimedoltomasbo: hi15:41
ltomasbonot sure I follow (or have the context)15:41
ltomasbothere is no ovs bridge inside the VM, right?15:41
ltomasboor you mean in the trunk ovs bridge?15:41
ltomasbowe don't need macvlan/ipvlan for the trunk port scenario (vlan)15:42
*** hongbin has joined #openstack-kuryr15:42
apuimedoI mean, instead of using ipvlan/macvlan/vlan on the VM15:45
apuimedowe create an ovs bridge there and use the internal ports15:45
apuimedonot sure how that can hit performance15:45
apuimedoit's probably worse for container-in-vm case15:45
apuimedofor baremetal it's probably better than using veth to br-int15:46
openstackgerritDarla Ahlert proposed openstack/kuryr-libnetwork: Unittest with mock  https://review.openstack.org/39708615:49
*** yamamoto has quit IRC15:49
apuimedovikasc: https://review.openstack.org/#/c/394324/215:52
ltomasboI think that for the trunk port, where there is already an extra ovs-trunk brdige15:52
ltomasbohaving yet another one inside the VM could be a bit too much, right? what would be the benefit of that (weighted by the complexity)?15:52
ltomasboon the other hand, for the baremetal case15:53
ltomasbowithout considering ipvlan/macvlan/vlan, it could be good to link the container directly to the OVS15:53
ltomasbowithout veth, same way as VMs are plugged with ovs-firewall15:54
apuimedoltomasbo: the benefit is having practically the same code as for baremetal15:54
apuimedogotta go for a while15:54
*** apuimedo is now known as apuimedo|away15:54
*** tonanhngo has joined #openstack-kuryr16:05
*** yamamoto has joined #openstack-kuryr16:06
*** tonanhngo has quit IRC16:07
*** jgriffith_away is now known as jgriffith16:09
*** dimak has quit IRC16:12
openstackgerritMerged openstack/kuryr-libnetwork: Unittest with mock  https://review.openstack.org/39708616:18
*** dougbtv has joined #openstack-kuryr16:20
*** tonanhngo has joined #openstack-kuryr16:32
*** tonanhngo has quit IRC16:33
janonymousalraddarla_: thanks for your work on rebasing16:42
janonymous https://review.openstack.org/#/q/status:open+project:openstack/kuryr-libnetwork+branch:master+topic:bp/drop-mox16:43
alraddarla_janonymous : I'm glad I could help :) Hoping to get more and more involved as I get more familiar with the project.16:44
*** tonanhngo has joined #openstack-kuryr16:45
*** tonanhngo has quit IRC16:46
openstackgerritMarco Chiappero proposed openstack/kuryr-libnetwork: Move interface creation and deletion  https://review.openstack.org/39454716:54
*** tonanhngo has joined #openstack-kuryr16:57
*** tonanhngo has quit IRC16:58
*** shashank_hegde has joined #openstack-kuryr17:06
*** tonanhngo has joined #openstack-kuryr17:10
*** tonanhngo has quit IRC17:13
*** yamamoto has quit IRC17:14
*** yamamoto has joined #openstack-kuryr17:15
*** yamamoto has quit IRC17:17
*** yamamoto has joined #openstack-kuryr17:18
*** jgriffith is now known as jgriffith_away17:19
*** shashank_hegde has quit IRC17:21
*** pcaruana has quit IRC17:23
*** tonanhngo has joined #openstack-kuryr17:26
*** tonanhngo has quit IRC17:27
*** lmdaly has quit IRC17:32
*** yamamoto has quit IRC17:34
openstackgerritIlya Chukhnakov proposed openstack/kuryr-kubernetes: [WIP] CNI & VIFBridge binding  https://review.openstack.org/40403817:35
*** garyloug_ has quit IRC17:44
*** yamamoto has joined #openstack-kuryr17:46
*** yamamoto has quit IRC17:51
*** jgriffith_away is now known as jgriffith18:00
*** dougbtv has quit IRC18:02
*** shashank_hegde has joined #openstack-kuryr18:10
*** dougbtv has joined #openstack-kuryr18:11
*** shashank_hegde has quit IRC18:26
*** oanson has quit IRC18:36
*** lezbar has quit IRC18:38
*** shashank_hegde has joined #openstack-kuryr18:50
*** tonanhngo has joined #openstack-kuryr19:08
*** tonanhngo has quit IRC19:12
*** saneax-_-|AFK is now known as saneax19:12
*** tonanhngo has joined #openstack-kuryr19:23
openstackgerritIlya Chukhnakov proposed openstack/kuryr-kubernetes: Experimental CNI & VIFBridge binding  https://review.openstack.org/40403820:13
*** saneax is now known as saneax-_-|AFK20:30
*** devlaps has quit IRC20:36
*** devlaps has joined #openstack-kuryr21:34
*** hongbin has quit IRC22:01
*** dougbtv has quit IRC22:50
*** pmannidi has joined #openstack-kuryr23:25
« Tuesday, 2016-11-29 Index Thursday, 2016-12-01 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!