I might or might not be workaround but only thing we as plugin should do is send the error message from switch to the controller. :) Anyway it would help to get the information from wireshark if plugin can help the issue with the switch.
If master role can be set the master role will be taken by another controller and this switch connection will be dropped. From: Anil Vishnoi [mailto:[email protected]] Sent: Thursday, April 6, 2017 9:54 PM To: Abhijit Kumbhare <[email protected]> Cc: Kochba, Alon <[email protected]>; Jozef Bacigál <[email protected]>; Srikanth Lingala <[email protected]>; [email protected]; odl netvirt dev <[email protected]>; Gorja Gorja <[email protected]>; Veera.Reddy B <[email protected]> Subject: Re: [netvirt-dev] [openflowplugin-dev] FW: br-int become slave and disconnect from ODL In my opinion, this is partially a controller issue and partially problem with the switch. The patches that jozef mention, it basically stop sending the slave role, but it doesn't mean that this issue won't occur when controller sends master role. First issue here is that switch is not accepting the role (it can be slave or master role), and when switch doesn't accept the role, controller just disconnect from the device and wait for the connection reattempt. Whatever you are seeing is expected behavior. Not sending slave role will workaround the issue, but what if controller sends master role and switch doesn't accept it ? You will end up in the same situation here. I would suggest to please collect the wireshark from the switch that is not accepting the role request and see what error message it's sending, that will help us to figure out whether it's an issue with ovs or controller should attempt anything else in case of failure. On Thu, Apr 6, 2017 at 10:14 AM, Abhijit Kumbhare <[email protected]<mailto:[email protected]>> wrote: The summary is here: https://wiki.opendaylight.org/view/OpenDaylight_OpenFlow_Plugin:Split_connection_layer Yes - it will be merged in the master first - we will get it in Boron only after it has been soaked well in Carbon first. Anil has been reviewing the set of patches since January - since then Shuva has reviewed them over the last weekend, followed by tests from Luis and now Prasanna is reviewing as one last check (he reviewed 8 of those yesterday & hopefully remaining will be done Friday). The tests with Luis have indicated things are better than before (in terms of number of failures in the tests) - but there are still 2 remaining cluster failures. Bug 6459 is kept open to track the 2 remaining cluster failures. We are targeting merge of the patches before the branch cutoff - but hopefully on Friday. On Thu, Apr 6, 2017 at 12:49 AM, Kochba, Alon <[email protected]<mailto:[email protected]>> wrote: Thanks guys. Could we have a summary of the patches and have them merged in master first before we start talking about Boron? What is this waiting on? Branch cutting is around the corner, we should allow some soaking time after merge to see there are no new issues. --alon From: [email protected]<mailto:[email protected]> [mailto:[email protected]<mailto:[email protected]>] On Behalf Of Jozef Bacigál Sent: Thursday, 6 April 2017 09:49 To: Abhijit Kumbhare <[email protected]<mailto:[email protected]>>; Srikanth Lingala <[email protected]<mailto:[email protected]>> Cc: odl netvirt dev <[email protected]<mailto:[email protected]>>; [email protected]<mailto:[email protected]>; [email protected]<mailto:[email protected]>; Gorja Gorja <[email protected]<mailto:[email protected]>>; Aizer, Koby <[email protected]<mailto:[email protected]>>; Veera.Reddy B <[email protected]<mailto:[email protected]>> Subject: Re: [openflowplugin-dev] [netvirt-dev] FW: br-int become slave and disconnect from ODL Let me check today if I am able to pick the fix from the changes and put in on the standing code in boron. It may or may not be possible. I will let you know. Jozef From: Abhijit Kumbhare [mailto:[email protected]] Sent: Thursday, April 6, 2017 7:45 AM To: Srikanth Lingala <[email protected]<mailto:[email protected]>> Cc: Aizer, Koby <[email protected]<mailto:[email protected]>>; Jozef Bacigál <[email protected]<mailto:[email protected]>>; odl netvirt dev <[email protected]<mailto:[email protected]>>; [email protected]<mailto:[email protected]>; [email protected]<mailto:[email protected]>; Veera.Reddy B <[email protected]<mailto:[email protected]>>; Gorja Gorja <[email protected]<mailto:[email protected]>> Subject: Re: [openflowplugin-dev] [netvirt-dev] FW: br-int become slave and disconnect from ODL I believe there are several patches by Jozef for the split connection handler. Since it is not really a new feature - rather a fix (albeit with a lot of code change) - it may be possible to backport to stable/boron but only Jozef can answer how easy it would be to backport as code may be depending on changes in Carbon. On Wed, Apr 5, 2017 at 10:23 PM, Srikanth Lingala <[email protected]<mailto:[email protected]>> wrote: Thanks for the info. I am also using stable/boron. Can I cherry pick these patches to stable/boron? Can you provide me the link to the patch file you are referring? Regards, Srikanth. From: Aizer, Koby [mailto:[email protected]<mailto:[email protected]>] Sent: Wednesday, April 05, 2017 10:53 PM To: Jozef Bacigál <[email protected]<mailto:[email protected]>>; Abhijit Kumbhare <[email protected]<mailto:[email protected]>> Cc: odl netvirt dev <[email protected]<mailto:[email protected]>>; Srikanth Lingala <[email protected]<mailto:[email protected]>>; [email protected]<mailto:[email protected]>; [email protected]<mailto:[email protected]> Subject: Re: [openflowplugin-dev] [netvirt-dev] FW: br-int become slave and disconnect from ODL Sounds great thanks. Can we get it cherry picked to stable/boron as well? We've been hitting this quite a lot after OVS restarts. Thanks, Koby ________________________________ From: Jozef Bacigál <[email protected]<mailto:[email protected]>> Sent: Apr 5, 2017 7:55 PM To: Abhijit Kumbhare; Aizer, Koby Cc: odl netvirt dev; Srikanth Lingala; [email protected]<mailto:[email protected]>; [email protected]<mailto:[email protected]> Subject: Re: [openflowplugin-dev] [netvirt-dev] FW: br-int become slave and disconnect from ODL Yes it should. Jozef ________________________________ Od: Abhijit Kumbhare <[email protected]<mailto:[email protected]>> Odoslané: 5. apríla 2017 18:43 Komu: Aizer, Koby Kópia: odl netvirt dev; Srikanth Lingala; [email protected]<mailto:[email protected]>; [email protected]<mailto:[email protected]> Predmet: Re: [openflowplugin-dev] [netvirt-dev] FW: br-int become slave and disconnect from ODL Jozef, Do you think this will be fixed when the split connection patches get merged? Abhijit On Wed, Apr 5, 2017 at 9:27 AM, Aizer, Koby <[email protected]<mailto:[email protected]>> wrote: + OFP Hi Srikanth, I think you're hitting [1], which I believe has a pending patch in master. Koby [1] https://bugs.opendaylight.org/show_bug.cgi?id=8087 ________________________________ From: Srikanth Lingala <[email protected]<mailto:[email protected]>> Sent: Apr 5, 2017 6:00 PM To: odl netvirt dev; [email protected]<mailto:[email protected]> Subject: [netvirt-dev] FW: br-int become slave and disconnect from ODL +netvirt-dev and ovs-discuss From: Srikanth Lingala Sent: Tuesday, April 04, 2017 11:11 PM To: '[email protected]<mailto:[email protected]>' <[email protected]<mailto:[email protected]>>; [email protected]<mailto:[email protected]> Subject: br-int become slave and disconnect from ODL Hi, I am trying to integrate Openstack with ODL. I am using ODL Boron stable release and OVS 2.6.1. I started ODL karaf using the below command: /opt/distribution-karaf-0.5.3-SNAPSHOT/bin# ./karaf -of13 clean Apache Karaf starting up. Press Enter to open the shell now... 100% [========================================================================] Karaf started in 273s. Bundle stats: 389 active, 401 total ________ ________ .__ .__ .__ __ \_____ \ ______ ____ ____ \______ \ _____ ___.__.| | |__| ____ | |___/ |_ / | \\____ \_/ __ \ / \ | | \\__ \< | || | | |/ ___\| | \ __\ / | \ |_> > ___/| | \| ` \/ __ \\___ || |_| / /_/ > Y \ | \_______ / __/ \___ >___| /_______ (____ / ____||____/__\___ /|___| /__| \/|__| \/ \/ \/ \/\/ /_____/ \/ Hit '<tab>' for a list of available commands and '[cmd] --help' for help on a specific command. Hit '<ctrl-d>' or type 'system:shutdown' or 'logout' to shutdown OpenDaylight. opendaylight-user@root> And, executed below command to set manager to ODL: #> ovs-vsctl set-manager tcp:10.20.0.26:6640<http://10.20.0.26:6640> ODL able to add br-int to OVS. Following are the OVS details: root@node-4446ubuntu:~# ovs-vsctl show 9bf5a6cb-e906-4205-afb2-14bb31510a51 Manager "tcp:10.20.0.26:6640<http://10.20.0.26:6640>" is_connected: true Bridge br-int Controller "tcp:10.20.0.26:6653<http://10.20.0.26:6653>" fail_mode: secure Port br-int Interface br-int type: internal Following are the ovs-vswitchd logs: 2017-04-04T22:30:46.005Z|01281|connmgr|INFO|br-int: added primary controller "tcp:10.20.0.26:6653<http://10.20.0.26:6653>" 2017-04-04T22:30:46.005Z|01282|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connecting... 2017-04-04T22:30:46.009Z|01283|netdev_linux|WARN|br-int: removing policing failed: Operation not supported 2017-04-04T22:30:46.576Z|01284|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connected 2017-04-04T22:31:51.646Z|01300|rconn|ERR|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: no response to inactivity probe after 5 seconds, disconnecting 2017-04-04T22:31:52.646Z|01301|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connecting... 2017-04-04T22:31:52.675Z|01302|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connected 2017-04-04T22:32:02.646Z|01303|rconn|ERR|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: no response to inactivity probe after 5 seconds, disconnecting 2017-04-04T22:32:03.645Z|01304|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connecting... 2017-04-04T22:32:03.665Z|01305|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connected 2017-04-04T22:32:13.646Z|01306|rconn|ERR|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: no response to inactivity probe after 5 seconds, disconnecting 2017-04-04T22:32:14.646Z|01307|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connecting... 2017-04-04T22:32:15.645Z|01308|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connection timed out 2017-04-04T22:32:15.646Z|01309|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: waiting 2 seconds before reconnect 2017-04-04T22:32:17.646Z|01310|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connecting... 2017-04-04T22:32:19.646Z|01311|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connection timed out 2017-04-04T22:32:19.647Z|01312|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: waiting 4 seconds before reconnect 2017-04-04T22:32:23.646Z|01313|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connecting... 2017-04-04T22:32:27.646Z|01314|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: connection timed out 2017-04-04T22:32:27.647Z|01315|rconn|INFO|br-int<->tcp:10.20.0.26:6653<http://10.20.0.26:6653>: continuing to retry connections in the background but suppressing further logging I am able to see the following ODL errors in the Karaf logs: 2017-04-05 04:01:08,675 | INFO | entLoopGroup-7-3 | ConnectionAdapterImpl | 363 - org.opendaylight.openflowjava.openflow-protocol-impl - 0.8.3.SNAPSHOT | Hello received / branch 2017-04-05 04:01:08,684 | INFO | entLoopGroup-7-3 | DeviceManagerImpl | 368 - org.opendaylight.openflowplugin.impl - 0.3.3.SNAPSHOT | ConnectionEvent: Device connected to controller, Device:/10.20.0.26:54319<http://10.20.0.26:54319>, NodeId:Uri [_value=openflow:167725479026029] 2017-04-05 04:01:08,685 | INFO | entLoopGroup-7-3 | SalRoleServiceImpl | 368 - org.opendaylight.openflowplugin.impl - 0.3.3.SNAPSHOT | SetRole called with input:SetRoleInput [_controllerRole=BECOMESLAVE, _node=NodeRef [_value=KeyedInstanceIdentifier{targetType=interface org.opendaylight.yang.gen.v1.urn.opendaylight.inventory.rev130819.nodes.Node, path=[org.opendaylight.yang.ge<http://org.opendaylight.yang.ge>n.v1.urn.opendaylight.inventory.rev130819.Nodes, org.opendaylight.yang.gen.v1.urn.opendaylight.inventory.rev130819.nodes.Node[key=NodeKey [_id=Uri [_value=openflow:167725479026029]]]]}], augmentation=[]] 2017-04-05 04:01:08,685 | INFO | entLoopGroup-7-3 | SalRoleServiceImpl | 368 - org.opendaylight.openflowplugin.impl - 0.3.3.SNAPSHOT | Requesting state change to BECOMESLAVE 2017-04-05 04:01:08,685 | INFO | entLoopGroup-7-3 | SalRoleServiceImpl | 368 - org.opendaylight.openflowplugin.impl - 0.3.3.SNAPSHOT | RoleChangeTask called on device:openflow:167725479026029 OFPRole:BECOMESLAVE 2017-04-05 04:01:08,685 | INFO | entLoopGroup-7-3 | RoleService | 368 - org.opendaylight.openflowplugin.impl - 0.3.3.SNAPSHOT | getGenerationIdFromDevice called for device: openflow:167725479026029 2017-04-05 04:01:18,695 | WARN | pool-36-thread-1 | DeviceContextImpl | 368 - org.opendaylight.openflowplugin.impl - 0.3.3.SNAPSHOT | New role BECOMESLAVE was not propagated to device openflow:167725479026029 during 10 sec 2017-04-05 04:01:18,695 | ERROR | pool-36-thread-1 | SalRoleServiceImpl | 368 - org.opendaylight.openflowplugin.impl - 0.3.3.SNAPSHOT | SetRoleService set Role BECOMESLAVE for Node: KeyedInstanceIdentifier{targetType=interface org.opendaylight.yang.gen.v1.urn.opendaylight.inventory.rev130819.nodes.Node, path=[org.opendaylight.yang.ge<http://org.opendaylight.yang.ge>n.v1.urn.opendaylight.inventory.rev130819.Nodes, org.opendaylight.yang.gen.v1.urn.opendaylight.inventory.rev130819.nodes.Node[key=NodeKey [_id=Uri [_value=openflow:167725479026029]]]]} fail . Reason java.util.concurrent.CancellationException: Task was cancelled. 2017-04-05 04:01:18,696 | WARN | pool-36-thread-1 | DeviceManagerImpl | 368 - org.opendaylight.openflowplugin.impl - 0.3.3.SNAPSHOT | Was not able to set role SLAVE to device on node openflow:167725479026029 Can anyone let me know what could be the issue? Regards, Srikanth. _______________________________________________ openflowplugin-dev mailing list [email protected]<mailto:[email protected]> https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev _______________________________________________ netvirt-dev mailing list [email protected]<mailto:[email protected]> https://lists.opendaylight.org/mailman/listinfo/netvirt-dev -- Thanks Anil
_______________________________________________ openflowplugin-dev mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev
