Hi Ryan, Thanks for forwarding but what is puzzling me is why the flow does not appear as it should in the Config Datastore. If the YANG model matches should it not appear in the config datastore?
I could see it as an openflow problem if I saw an exception why trying to instantiate the flow. I was suspecting MDSAL for this reason. Perhaps I am mistaken. Thanks. Ranga On Wed, Dec 6, 2017 at 10:42 AM, Ryan Goulding <[email protected]> wrote: > This is an OpenFlow issue; forwarding your request to openflowplugin-dev. > > Kind Regards, > > Ryan Goulding > > On Wed, Dec 6, 2017 at 10:40 AM, M. Ranganathan <[email protected]> wrote: > >> What is interesting is that a PUSH and output WITHOUT the GOTO does >> appear. Why does the GOTO cause the problem? >> >> Clearly this is a legal flow -- I can construct it with ovs-ofctl but not >> with opendaylight. >> >> What am I doing wrong? >> >> On Tue, Dec 5, 2017 at 4:22 PM, M. Ranganathan <[email protected]> wrote: >> >>> >>> On Tue, Dec 5, 2017 at 2:14 AM, Tom Pantelis <[email protected]> >>> wrote: >>> >>>> >>>> >>>> On Mon, Dec 4, 2017 at 6:42 PM, M. Ranganathan <[email protected]> >>>> wrote: >>>> >>>>> >>>>> Hello, >>>>> >>>>> I am working with the openflow southbound trying to install a flow >>>>> that installs a vlan push action with a goto and an output. >>>>> >>>>> I believe I have hit an MDSAL issue because all I see in the config >>>>> data store is: >>>>> >>>>> >>>>> First, is this a known issue for the Carbon release that has been fixed >>>>> >>>>> (is there a bug tracker for MDSAL where I can check?) If so I'll move up >>>>> to Nitrogen. >>>>> >>>>> Second, how can I debug MDSAL? I'd like to selectively turn on debug >>>>> tracing for that component only. >>>>> >>>>> How can I set up debugging ONLY for that component so I can figure >>>>> >>>>> out what is going on? >>>>> >>>>> >>>> I'm not aware of any issue in the data store where a client submits >>>> data but it somehow doesn't get written/applied or is otherwise missing >>>> when retrieved. It will return an error (and log it) if it can't commit the >>>> transaction for any reason. >>>> >>>> To see transaction activity on the front-end side, you can enable this >>>> debug: >>>> >>>> log4j.logger.org.opendaylight.controller.cluster.datastore. >>>> TransactionProxy=DEBUG >>>> >>>> This will log all writes, merges, and deletes so you can see if you're >>>> expected data was actually written (or if it got deleted). >>>> >>> >>> >>> I did what you suggested and logged the instructions from the flow just >>> before commit. I turned on debugging. The commit succeded (no exceptions). >>> So the flow appears in the config but no Vlan push or set instruction or >>> output only the goto table instruction. It appears that something stripped >>> these off in the process of writing to config datastore. >>> The match here is on metadata. >>> >>> In the logs, I simply see : >>> >>> | 288 - gov.nist.sdnmud.impl - 0.1.0.SNAPSHOT | >>> org.opendaylight.yang.gen.v1.urn.opendaylight.flow.inventory >>> .rev130819.tables.table.FlowBuilder@470b2bcb >>> 2017-12-06 02:48:35,843 | DEBUG | tp1317035402-582 | >>> TransactionProxy | 204 - >>> org.opendaylight.controller.sal-distributed-datastore >>> - 1.5.2.Carbon | New READ_WRITE Tx - member-1-datastore-config-fe-1 >>> -txn-121-0 >>> 2017-12-06 02:48:35,843 | DEBUG | tp1317035402-582 | >>> TransactionProxy | 204 - >>> org.opendaylight.controller.sal-distributed-datastore >>> - 1.5.2.Carbon | Tx member-1-datastore-config-fe-1-txn-121-0 >>> executeModification MergeModification /(urn:opendaylight:inventory?r >>> evision=2013-08-19)nodes >>> 2017-12-06 02:48:35,843 | DEBUG | tp1317035402-582 | >>> TransactionProxy | 204 - >>> org.opendaylight.controller.sal-distributed-datastore >>> - 1.5.2.Carbon | Tx member-1-datastore-config-fe-1-txn-121-0 >>> executeModification MergeModification /(urn:opendaylight:inventory?r >>> evision=2013-08-19)nodes/node >>> >>> >>> But not the actual data that was written to the data store. What other >>> logging should I enable to see this? >>> >>> >>> I can provide further details or create a test case abstracting the >>> problem. Thanks, >>> >>> >>> >>> >>> >>> >>> >>> >>>> >>>> >>>>> >>>>> I would much appreciate some help getting started and I am willing to >>>>> volunteer my time and >>>>> >>>>> effort to fixing bugs - should there be one. >>>>> >>>>> Thank you for your help in this matter >>>>> >>>>> Regards, >>>>> >>>>> >>>>> Ranga >>>>> >>>>> >>>>> -- >>>>> M. Ranganathan >>>>> >>>>> “If debugging is the process of removing software bugs, then >>>>> programming must be the process of putting them in.” – Edsger Dijkstra >>>>> >>>>> >>>>> _______________________________________________ >>>>> mdsal-dev mailing list >>>>> [email protected] >>>>> https://lists.opendaylight.org/mailman/listinfo/mdsal-dev >>>>> >>>>> >>>> >>> >>> >>> -- >>> M. Ranganathan >>> >>> “If debugging is the process of removing software bugs, then programming >>> must be the process of putting them in.” – Edsger Dijkstra >>> >>> >> >> >> -- >> M. Ranganathan >> >> “If debugging is the process of removing software bugs, then programming >> must be the process of putting them in.” – Edsger Dijkstra >> >> >> _______________________________________________ >> mdsal-dev mailing list >> [email protected] >> https://lists.opendaylight.org/mailman/listinfo/mdsal-dev >> >> > -- M. Ranganathan “If debugging is the process of removing software bugs, then programming must be the process of putting them in.” – Edsger Dijkstra
_______________________________________________ openflowplugin-dev mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev
